E-Mobility as a Sustainable System Innovation

Transcription

1 E-Mobility as a Sustainable System Innovation Insights from a Captured Niche Dissertation Zur Erlangung des Grades Doktor der Wirtschaftswissenschaft (Dr. rer. oec.) des Fachbereich B Wirtschaftswissenschaft Schumpeter School of Business and Economics an der Bergischen Universität Wuppertal vorgelegt von Karoline Augenstein Wuppertal 2014

2 Datum der Einreichung: Erstgutachter: Zweitgutachter: Prof. Dr. Uwe Schneidewind Prof. Dr. Ulrich Braukmann Datum der Prüfung:

3 Schriftenreihe der Reiner Lemoine-Stiftung Karoline Augenstein E-Mobility as a Sustainable System Innovation Insights from a Captured Niche Shaker Verlag Aachen 2015

4 Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at Zugl.: Wuppertal, Univ., Diss., 2014 Copyright Shaker Verlag 2015 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the publishers. Printed in Germany. ISBN ISSN Shaker Verlag GmbH P.O. BOX D Aachen Phone: 0049/2407/ Telefax: 0049/2407/ Internet: info@shaker.de

5 OVERVIEW OF CONTENT Overview of Content... i Table of Contents... iii List of Figures... vii List of Tables... vii List of Abbreviations... viii Abstract... ix 1 Introduction Research Questions Research Approach and Contribution of the Thesis Structure of the Thesis E-Mobility as a Sustainable System Innovation Definitions Visions and Concepts of Sustainable Mobility E-mobility and the Energy Transition E-mobility as a Sustainable System Innovation 3 Traps Current State of Research: Literature Review A Theoretical Framework for Analyzing System Innovation Sustainability Transitions Research Analyzing Transitions in the Making Developing a Theoretical Framework for Analyzing System Innovations E-mobility as a Sustainable System Innovation the Case of Germany Methodology Delineating the System Innovation as an Ideal-Type The Transformative Capacity of the Electric Vehicle Case Study: E-mobility in Germany Assessing System Adaptability i

6 4.5 E-mobility and the Potential for Sustainable System Innovation: 3 Traps Revisited Conclusions and Outlook References List of interviews Appendix Danksagung ii

7 TABLE OF CONTENTS Overview of Content... i Table of Contents... iii List of Figures... vii List of Tables... vii List of Abbreviations... viii Abstract... ix 1 Introduction Research Questions Research Approach and Contribution of the Thesis Structure of the Thesis E-Mobility as a Sustainable System Innovation Definitions Visions and Concepts of Sustainable Mobility E-mobility and the Energy Transition E-mobility as a Sustainable System Innovation 3 Traps Current State of Research: Literature Review A Theoretical Framework for Analyzing System Innovation Sustainability Transitions Research Central Concepts and their Origins The Role of Technology and Innovation The Multi-Level Perspective on Sustainability Transitions (MLP) The Concept of Regime Structuration and the Potential for Change Transition Patterns and Pathways Niche-Regime Dynamics Multi-Regime Dynamics Critical Reflection: Shortcomings and Limitations Analyzing Transitions in the Making iii

8 3.2.1 Strategic Niche Management (SNM) Transition Management (TM) Fostering Transitions in the Making through SNM and TM? Transitions in the Making : How to Analyze the Potential for System Innovation ex ante? Developing a Theoretical Framework for Analyzing System Innovations Unit of Analysis: Socio-technical System and Regime Socio-technical Change and the Concept of System Innovation The Transformative Capacity of New Technologies System Adaptability An Integrated Analytical Framework: Synthesis and Discussion E-mobility as a Sustainable System Innovation the Case of Germany Methodology Delineating the System Innovation as an Ideal-Type E-mobility and Renewable Energy E-mobility in Fleets E-mobility and Intermodality E-mobility and Carsharing The Transformative Capacity of the Electric Vehicle Case Study: E-mobility in Germany Assessing System Adaptability Case Study Design The German Innovation System for E-mobility Political Initiatives by the Federal Government: Caught between Environmental and Economic Goals Actor Structure in the German Innovation System for E-mobility: A Key Role for Regime Incumbents Funding Programs: Trying to Achieve Systemic Change through Technological Innovation Baden-Württemberg The Automotive State iv

9 E-mobility Initiatives in Baden Württemberg Model Region Stuttgart Leading-Edge Cluster Electric Mobility South-West Showcase Region LivingLab BW e mobil Other Initiatives Network Analysis Dominant Actors and Cooperation Patterns for Developing E-mobility in Baden-Württemberg Identifying Actors in the Network Identifying Relations and Power Positions in the Network Identifying Patterns of Cooperation Patterns of Cooperation in the Model Region Region Stuttgart : Modest Attempts at Cross-Sector Cooperation in a Traditional R&D Setting Patterns of Cooperation in the Showcase Region Living Lab BW e mobil : New Actors in Diverse Cooperation Settings and Disillusioned Regime Incumbents Patterns of Cooperation in the Leading-Edge Cluster Electric Mobility South-West : Focusing on Industrial R&D for E-mobility in the Automotive Sector Clustering Projects: Limited Potential of Multi-Regime Dynamics and Unexpected Seeds of System Innovation Projects focusing on vehicle technology: Cross-sectoral cooperation patterns without systemic perspectives Projects focusing on systemic e-mobility: Broadening perspectives outside cross-sectoral cooperation settings Projects focusing on system-innovative e-mobility: New impulses from ICT, mobility services and actors entering the field of e-mobility from outside transport or energy regimes Discourses and Dominant Frames of Reference: System Adaptability for Sustainable E-mobility Synthesis: Shared Hopes for the Future Varied Perceptions of Current Problems and Suitable Strategies A Closer Look at System-Innovative E-mobility Projects in Baden-Württemberg v

10 Changing Frames of Reference: High Hopes for Future E-mobility based on Diverging Intentions and Problem Perceptions Changes in the Technological Profile of the Socio-technical System: Risk and Uncertainty in a Transitional Phase Drivers and Barriers for E-mobility: The Transformative Capacity of Public Funding for E-mobility New Forms of Cooperation for E-mobility: The Important Role of a Neutral Coordination Agency in a Situation of Limited Commitment and Avoided Conflicts New Actors Entering the Field of E-mobility: Potential for System Innovation through Fresh Perspectives and Resourceful Outsiders Discussion: Future Visions, Transformation and Cooperation in Developing E- mobility Synthesis: Lacking Concepts for Dealing with System Innovation and Processes of Transformation in Practice Baden-Württemberg and the German Innovation System for E-mobility: A Captured Niche? E-mobility and the Potential for Sustainable System Innovation: 3 Traps Revisited The Quantitative Trap : Purely Market-Based Diffusion of Electric Vehicles is not Feasible The Qualitative Trap : Advancing System Adaptability for E-mobility through Outsiders and Coordinating Actors The Sustainability Trap : Avoiding Conflicts Synthesis: A Need for Experiments and Learning in Developing E-mobility as a Sustainable System Innovation Conclusions and Outlook References List of interviews Appendix Danksagung vi

11 LIST OF FIGURES Fig. 1: Battery-electric vehicles from (BMVBS, 2011, p. 5)... 3 Fig. 2: Structure of the Thesis - Main chapters and course of investigation Fig. 3: Mutual dependency and interlinkages in the field of energy and mobility transitions. 32 Fig. 4: Three traps of a sustainable system innovation Fig. 5: Distribution of papers across fields of research Fig. 6: Topical focus of papers Fig. 7: Distribution of selected papers across three traps Fig. 8: Multi-level perspective on transitions (based on Geels & Schot, 2010, p. 25) Fig. 9: Technology-based sectoral change: basic categories (Dolata, 2009, p. 1067) Fig. 10: An integrated framework for analyzing system innovations Fig. 11: Case study design Fig. 12: Electric Mobility in Model Regions Overview (Sax-Mobility II, 2012) Fig. 13: Showcase Regions for Electric Mobility Overview (GGEMO, 2013) LIST OF TABLES Tbl. 1: Overview of papers with a focus on e-mobility as a sustainable system innovation.45 Tbl. 2: Typology of transition pathways, based on: Geels & Schot, 2007, p. 406 ff Tbl. 3: The potential for system innovation: transformative capacity and system adaptability Tbl. 4: The transformative capacity of the battery-electric vehicle Tbl. 5: Project funding in model regions and share of funding for system-innovative projects (data based on BMBF 2011, Tenkhoff et al., 2011) Tbl. 6: Overview of actors in the innovation system for e-mobility in Baden-Württemberg and their sectoral affiliations Tbl. 7: Levels of influence in the innovation system network: Concentrated power and dominant regime actors Tbl. 8: Patterns of cooperation in the model region Region Stuttgart based on affiliation and influence of project partners vii

12 LIST OF ABBREVIATIONS BEV Battery-Electric Vehicle BMBF German Federal Ministry of Education and Research BMU German Federal Ministry for the Environment, Nature Conservation and Nuclear Safety BMUB German Federal Ministry for the Environment, Nature Conservation, Building and Nuclear Safety (former BMU) BMVBS German Federal Ministry of Transport, Building and Urban Development BMVI German Federal Ministry of Transport and Digital Infrastructures (former BMVBS) BMWi German Federal Ministry of Economics and Technology CHP Combined Heat and Power e-mobil BW State Agency for Electric Mobility and Fuel Cell Technology e-mobil BW GmbH FCV Fuel Cell Vehicle GGEMO Joint Agency for Electric Mobility of the German Federal Government GHG Greenhouse Gases ICE Internal Combustion Engine ICT Information and Communication Technology IEKP Integrated Energy and Climate Program of the German Federal Government MLP Multi-Level Perspective NEPE National Development Plan for Electric Mobility of the German Federal Government NOW GmbH National Organization for Hydrogen and Fuel Cell Technology NPE National Platform Electric Mobility OEM Original Equipment Manufacturer OPEC Organization of the Petroleum Exporting Countries PLS Regional Project Coordination Agency SME Small and Medium Sized Enterprises SNM Strategic Niche Management TIS Technological Innovation System TM Transition Management V2G Vehicle-to-Grid WRS Stuttgart Region Economic Development Corporation viii

13 ABSTRACT In 2009, the German government launched its National Development Plan for Electric Mobility which, in an attempt to integrate economic and environmental strategies, aims at developing Germany as a lead market for electric mobility and set the concrete target of having 1 million electric vehicles on the road by However, there have been hypes around e-mobility before and even if this goal were to be reached, a merely quantitative aim of a certain number of electric vehicles will not suffice to contribute to a more sustainable development in transport. This requires a more comprehensive vision of sustainable e- mobility including the development of renewable energies as electricity source, reduction of the volume of privately owned cars and new mobility patterns, i.e. a system innovation. Thus, the question addressed in this thesis is: How can we assess at this critical early stage whether there is potential for e-mobility developing as a sustainable system innovation? A theoretical framework will be developed for assessing the potential of a wider transition at an early stage by analyzing current patterns of socio-technical co-evolution and embedding these in a wider framework of the structural dynamics involved in transitions. It will be applied to the case of the German region Baden-Württemberg, where two of the largest demonstration projects for e-mobility are taking place. The aim of the analysis is to identify whether system-innovative projects do emerge in this specific case and what patterns (e.g. in terms of specific actor constellations, institutional adjustments etc.) can explain this. It will be shown that the system-innovative potential of this e-mobility niche remains limited, due to the powerful influence of incumbents, conflicting political goals and traditional science approaches. A few more system-innovative activities emerge where powerful actors from outside are involved, who are capable of viewing mobility in a more systemic way (e.g. actors from the public transport or housing sector). It is argued that the role of large demonstration projects is important, but they need to be designed as transdisciplinary research projects from the beginning. ix

14 1 INTRODUCTION The transport sector currently represents one of the key challenges for future sustainable development, while at the same time efforts of making mobility more sustainable have an enormous potential for more sustainable societies. Personal mobility is a central feature of modern societies; the freedom and the affordability of the means to move around are highly valued achievements with many positive effects on personal and social well-being. Global passenger and freight traffic has brought the world closer together. Simultaneously, the transport sector has emerged as a heavy burden on the natural environment. Almost a third of total global energy consumption can be allocated to the transport sector (IEA, 2013, p. 246) and more than 90% of this energy is generated from fossil sources, oil in particular (p. 510). This directly links the transport sector to the more basic challenge of an energy transition, breaking out of carbon lock-in (Unruh, 2000) and taking a new path based on renewable energy sources. The fact that in Europe more than half of all CO 2 -emissions are caused by transport, electricity and heat generation (EEA, 2013, p. 14 f.) shows that an integrated approach to an energy transition (including the transport and the energy sector) could have a substantial environmental effect and would possibly even benefit from synergies and spillovers. The pressure exerted by increasingly strict climate and environmental regulation, especially regarding CO 2 -emissions, has led to at least some concern and effort towards sustainability-oriented change in the transport sector. Typically, this type of change is hoped to be achieved with the help of technology-based innovations such as the battery-electric vehicle (BEV) as a technological fix that combines the reduction of emissions and the achievement of economic success with an innovative vehicle while preserving established mobility patterns. Additionally, in the case of e-mobility there is a prospect of integrating energy and transport transitions in mutually beneficial ways. Thus, it is hoped that the electric vehicle, a not so new innovation in automobility, will be the boundary object connecting (and solving) economic and environmental problems in the way we provide our societies with energy and mobility. Against this background, the German government launched its National Development Plan for Electric Mobility in The aim is to integrate economic and environmental strategies, in order to develop Germany as a lead market for electric mobility. A somewhat symbolic target has been set of having one million electric vehicles on the road by However, a brief historic overview shows that there have been hypes around e-mobility 1

15 before and it seems questionable whether the current one will be the exception and manage to persist. Electric vehicles were invented already during the 19 th century in the early days of the automobile age. A commonly referred to milestone for the automobile age is the appearance of Carl Benz first motorcar, which was demonstrated publicly in 1885 for the first time. At that time, alternative drive technologies were developed simultaneously, i.e. electric, steam engine and internal combustion engine (ICE), and all competed for market success. Especially battery-electric cars were common in large European capital cities and in New York as part of commercial fleets, e.g. taxi, garbage collection, postal or delivery services. While the internal combustion engine car began to force electric vehicles out of the market over the course of the 20 th century, they were temporarily revived during the two World Wars. This was mainly due to gasoline shortages, since ICE vehicles and fuels were primarily used for military purposes. However, already before the outbreak of World War II, the production of electric vehicles had almost ceased to exist during the economic crisis in the late 1920s. Until the mid-20 th century, electric vehicles had more or less disappeared from the scene (Høyer, 2008, p. 63 ff.). When public awareness for environmental problems rose from the mid-1960s onwards, the electric car was beginning to re-emerge as a potential solution especially to local air pollution and resource scarcity. The oil crisis in the early 1970s stimulated debates about the finite nature of fossil resources and the OPEC oil embargo showed that this could have very real consequences, e.g. the introduction of car-free Sundays and speed limits in Germany in However, early attempts of re-developing electric vehicles during the 1970s did not achieve any visible momentum apart from various R&D efforts that eventually failed to produce marketable results (Canzler & Knie, 2011, p. 102; Høyer, 2008, p. 66). Still, the electric vehicle did not completely vanish from the agenda, because environmental problems persisted and were increasingly discussed in a more global perspective and in the broader terms of sustainability and energy transition which links up with the question of alternative fuels and drive technology. At the same time, local problems of air pollution in metropolitan areas also aggravated, a prominent example being Los Angeles, and the Californian zero-emission regulations of 1990 are commonly referred to as an important milestone in the resurfacing of the electric vehicle during the 1990s (Canzler & Knie, 2011, p. 102 f.; Høyer, 2008, p. 67). It spurred a phase of more serious R&D efforts and in Europe large demonstration projects for e-mobility were carried out in Germany, France and Switzerland. These were hoped to lead to market breakthrough for electric vehicles, which however did not materialize (Hoogma et al., 2002, p. 67 ff.; Weider et al., 2011b, p. 108). 2

16 Even though the circumstances as well as the motivation for developing electric vehicles have differed across the past 150 years, the reasons why e-mobility has failed time and again have not changed substantially. Compared to conventional ICE cars, electric vehicles have always performed worse as regards range, price and charging time. Since this has held true for a considerable amount of time and in spite of various attempts to politically support the introduction of electric vehicles, it is reasonable to ask whether these are necessary structural limitations, and whether the electric car can really become more than a niche vehicle for some limited urban use purpose (Høyer, 2008, p. 71)? And indeed, when analyzing the potential of e-mobility as a solution for today s sustainability challenges in the transport sector, the perspectives according to industry- and technology-oriented discourses seem rather bleak even though there have been increased efforts to foster the development of electric vehicles since roughly Not long after the last phase of hype and eventual disappointment around BEVs during the 1990s, there is now a renewed interest, due to an increasing awareness of problems such as climate change and peak oil and their implications for policy and prices. However, why has the current phase in developing e-mobility already persisted for a period of roughly ten years, when there is broad skepticism, no real breakthrough, and a rich history of trial followed regularly by error? Fig. 1: Battery-electric vehicles from (BMVBS, 2011, p. 5) A number of factors and their simultaneous occurrence and interlinkages may play a role here. A window of opportunity has emerged in the intensified debates about climate change and research scarcity that have reached a peak in public and political attention around the mid-2000s. E-mobility re-appeared as a possible solution because it promises emission-free mobility and independence from oil (and oil producing countries). Due to a broadening and more integrated perspective on sustainability, as compared to the earlier sustainability discourse of the 1980s, e-mobility offered additional advantages. It would not produce some of the problem shifts, for instance those associated with bio-fuels that threaten to cause a 3

17 competition with food production for arable land (cf. Bringezu et al., 2009; 2012). Furthermore, e-mobility was also increasingly seen as a central element, or even a catalyst, in the context of the transition to renewable energies (Canzler & Knie, 2011, p. 109 f.). However, environmental concerns alone have so far not been decisive for developing e- mobility, and they probably will not be this time. A difference may be detected by looking at a number of additional factors surfacing during the mid-2000s that set this period apart from earlier phases of developing e-mobility. For one thing, the global economic and financial crisis severely hit the automotive industry (Schwedes, 2011, p. 13). In combination with environmental concerns, these economic difficulties created pressure not only for the industry, but especially also for politicians to search for alternatives. Since at that point electrification seems to be the more viable technological option for the future as compared to bio-fuels or hydrogen fuel cell technology governments have decided to invest in research and development of this technological field. Apart from the general recession, many car manufacturers are faced with saturated markets (at least in the industrialized countries) and thus attempt to diversify their strategies (Orsato et al., 2012, p. 210; Orsato & Wells, 2007, p. 999). Incentives for public or private investment in research and development of BEVs are further created by the improvements in battery technology achieved during that phase. Even though the basic characteristics and technological components of electric vehicles have been invented more than 100 years ago, those recent leaps in battery performance have caused hopes, not of radically different and new opportunities, but of finally coming closer to solving the age-old problem of a BEV s limited range and gradually developing electric vehicles that can compete with conventional ICE cars (Schwedes, 2011, p. 14). In accordance with the observed combination of environmental and economic concerns as well as geopolitical strategies, e.g. in relation to oil-producing countries and emerging economies (especially China), government spending on developing e-mobility is usually part of economic recovery plans, especially in countries where the automotive industry is central for the national economy. Germany is a typical case in this respect. It has an important automotive industry and the government has launched a comprehensive funding program around e-mobility for, mainly, economic reasons, but also in the context of its sustainability-related efforts. So far, these hopes have not materialized. In the German case, the point of market breakthrough for BEVs and substantial diffusion of innovative e-mobility concepts seems far away and there are major uncertainties with regard to the success of supporting measures and efforts altogether. Apart from regulatory and infrastructure issues, a major obstacle is 4

18 perceived to be the technological shortcoming of the battery as drive technology with regard to its range and price. Thus, despite some serious efforts, there are no signs of breakthrough on a broader scale and electric vehicles remain confined to small (market) niches and often publicly funded demonstration projects. Pilot and demonstration projects have emerged as the method of choice when it comes to investing in R&D for e-mobility (next to economic incentives for consumers to reach higher diffusion rates). The problem with demonstration projects is that due to their important function as a protected space, they usually amount to nothing more than a temporary meeting ground (Schot et al., 1994, p. 1073). They are quickly being dissolved again once the circumstances change, such as e.g. falling oil prices once the oil crisis had been overcome in the 1970s, and commitment by industry or public investors is lost (ibid.). Nonetheless, the role of pilot and demonstration projects can be crucial, especially if they are designed as experiments that focus not only on technological improvement but also social learning processes. For instance, it can be shown that dissatisfaction with the limited range of electric vehicles decreases among test users, because practical experience shows that a BEV is sufficient for most trips in everyday life (Weider et al., 2011b, p. 111). Processes of social learning are also important because from a sustainability perspective a purely technological solution, in this case electric vehicles substituting for conventional ICE cars, will not suffice to meaningfully address the complex challenges the transport sector is facing. Adopting a broad conceptualization of sustainability, transport-related problems include ecological issues such as greenhouse gas emissions and resource use as well as congestion in urban areas, issues of land use and spatial planning, and social issues of access to mobility and social equity. Assuming that at least part of the goal of developing e-mobility is to contribute to sustainable development, then the question is whether a more comprehensive perspective can be aligned with a technological perspective on electric vehicles? What is clear at least is that even if improvements in battery technology result in competitive BEVs with similar range and prices as conventional cars, a merely quantitative aim of achieving specific diffusion rates and numbers of electric vehicles on the market will not contribute significantly to sustainable development. What is needed is a more comprehensive vision of sustainable e-mobility, which includes technological as well as social change, e.g. in terms of new mobility patterns, infrastructures, policies and business models, and an integration with the development of renewable energies. Such a complex and fundamental change process can be described in terms of a system innovation. The specific strength of experiments shielded from market 5

19 pressure is that they can in principle facilitate learning processes including these various technological and social aspects, much rather than for instance economic incentives for consumers to buy electric vehicles. Some developments indicate that the circumstances for e-mobility developing as a system innovation contributing to sustainable development are favorable at the moment, as opposed to earlier attempts in the past decades. First of all, it is noteworthy that the term emobility has gained prominence in the relevant discourses. This indicates that the perspective has shifted beyond isolated views on the battery-electric vehicle itself. This was common during earlier periods where the focus was more exclusively on developing alternative drivetrains and vehicles (Canzler & Knie, 2011, p. 102.). Second, a broadening of perspective can also be observed with regard to the interlinkages of debates about e-mobility and the energy transition to renewable energies. While during the 1990s the electrification of cars was a project clearly confined within transport policy and the automotive industry, this has currently changed to some degree. Environmental problems in the transport sector are increasingly framed as part of the wider energy transition challenge. E-mobility thus becomes one of the strategies fostering the transition from a carbon-based to a renewables-based society (Weider & Rammler, 2011, p. 6). In practice, this means that developing e-mobility is no longer an issue only in the automotive industry but also includes new actors, especially from the energy sector (Canzler & Knie, 2011, p. 103). This implies that a comprehensive perspective on sustainable mobility can well be aligned with the development of electric vehicles. In fact, it is not an artificial burden (analytically or in practice) to focus on issues of broader sustainability when dealing with e- mobility, a broadened perspective may much rather be a new context for electric vehicles, where it is more likely that they will not disappear after a short hype, because they might evolve as a central technology in a new energy system. This is very different from a situation where the electric vehicle is developed as a perfect substitute for conventional cars. Thus, such a broad reframing in terms of electric mobility (instead of vehicle ) may help shaping better strategies for improving the environmental performance of the transport sector and may also become crucial for achieving an overall energy transition. Of course, it is highly uncertain whether such a potential can actually be realized and may be compared to the early days of the automobile age where cars were in the beginning referred to as horseless carriages (Høyer, 2008, p. 64). Reflecting on the massive consequences that the diffusion of the automobile had in the following and how it shaped many aspects of modern societies, it is clear that what can be observed currently is at most a transitional phase that is only beginning 6

20 to take shape, where it is still a major challenge to begin to re-think (auto)mobility in general. Against the background of its turbulent history, this implies that the electric car can be characterized as a radical innovation and, in principle, this has two possible consequences: the electric car may either trigger the kind of paradigmatic change characterizing sustainability transitions (or the emergence of automobility at the beginning of the past century); or it may fail altogether and never be able to become more than a technological niche, as has already been the case in history several times before. 1.1 Research Questions Since the current momentum in the electrification of the car, together with the emergence of broader concepts of e-mobility, seem to fuel hope for a transition in mobility, the question addressed in this thesis is: How can we assess at this critical early stage whether there is potential for e-mobility developing as a sustainable system innovation? What can be observed is an old innovation, the electric vehicle, that has failed several times before and is now being rediscovered, facing familiar problems (limited range, long charging times, high price) but also promising new opportunities and synergies in the context of interlinked sustainability challenges in the fields of mobility and energy. The scope of the research question thus goes beyond an issue of diffusion of a technological innovation; here, the concept of system innovation is central. The research thus builds on a long-standing tradition of economic scholarship on innovation, looking beyond specific markets or individual firm profit, but rather aims at understanding the way that novelty enters the world, the emergence of groundbreaking change and how technological innovation is embedded in and interrelated with its social context. These basic lines of thought have been an important influence on innovation scholars of different denominations studying the potential for change and innovation in the emerging field of sustainability transitions research. Since sustainability is a contested and ambiguous concept, involving large degrees of uncertainty, it is important to be specific on the way it is defined for the purpose of this thesis, in order to guarantee transparency and avoid normative distortions of the research. It is thus asked: 7

21 What is e-mobility, when understood as a sustainable system innovation? What are the different dimensions of such a concept? How can a sustainability dimension be included in a meaningful way? Based on these clarifications regarding the concept of e-mobility as a sustainable system innovation in theory, the central motivation for this thesis is finding a way of assessing whether current developments in the field of e-mobility may or may not indicate a potential towards the emergence of a system innovation. This question is relevant for many other empirical cases in the field of sustainability transitions research, focusing on technological innovations, such as alternative drivetrains or renewable energy technologies. The theoretical framework developed in this thesis aims at a better understanding of how technological innovations and social practices interrelate and how it can be assessed early on whether emerging system innovations are producing sustainable outcomes. How can potential for system innovations to emerge be assessed ex ante? Such an assessment will have to account for case-specific conditions and influencing factors and the theoretical framework will be explored in a case study of developing e-mobility in Germany. For this specific case it is asked: What are forces and barriers of a sustainable system innovation developing around e- mobility? How do different types of actors involved in developing e-mobility and their specific strategies foster or hinder the potential for system innovation? Since e-mobility crosses sectoral boundaries (especially automotive and energy), what types of cooperation patterns play a role? What are relevant framework conditions? In how far do public R&D funding programs and demonstration projects contribute to the development of system-innovative potential? In order to address these questions, a theoretical approach is needed that transcends the logic of incremental innovation and clearly conceptualizes processes of systemic socio-technical change. 8

22 1.2 Research Approach and Contribution of the Thesis This thesis is positioned as a contribution to the field of sustainability transitions research. Transitions are long-term and fundamental change processes in socio-technical systems and they are the core of this young field of research, which has emerged from various disciplinary backgrounds, and focuses on innovation and transformation processes towards more sustainable energy and transport systems (as well as other relevant societal domains in the context of sustainable development, such as water or agri-food systems). Transitions in these societal sub-systems are captured in terms of system innovation, i.e. the interplay of technological innovations and changes in social practices and institutions. Processes of sociotechnical co-evolution lead to radical and paradigmatic change in the way that specific societal functions, e.g. mobility or energy provision, are being fulfilled. The emergence of such new functionalities depends not only on technological innovation, or only on behavioral change. Neither will it be possible to introduce system innovation in a top-down fashion through command-and-control policies or straightforward management, just like bottom-up initiatives alone will in most cases fail to affect structural change. The concept of system innovation thus aims at capturing the interplay of technological and social innovation across different levels and interlinked sub-systems involved in fulfilling certain societal functions, such as energy or transport. For an analysis of whether and how e-mobility may contribute to a more sustainable transport system, i.e. in terms of reducing emissions and resource intensity as well as addressing issues of congestion, socially just access to mobility and interlinkages with the transition to renewable energies, system innovation is thus a useful concept addressing the inherent complexities. The research approach of this thesis, and most of transition studies in general, is therefore decidedly interdisciplinary, exploratory and based on broad heuristic analytical frameworks. This can be challenging with regard to methodology but the aim is to close a long existing research gap in more traditional scientific approaches of studying sustainability problems in the transport sector. Traditionally, there is a sharp separation between research on technological innovation and research on behavioral or institutional change as two completely separated means of addressing environmental problems caused by transport (Lyons, 2012, p. 30). This has to do with the fact that [t]ransport research on climate change mitigation tends to revolve around the reduction of carbon use given existing economic, social and political systems and ideals (Schwanen et al., 2011, p. 1004). Within such given structural conditions, disciplinary science often focuses on clearly delineated elements of such systems in separate strands of research. Most pronounced is the separation between a technological 9

23 perspective on emissions reduction through alternative drivetrains or fuels, and a social perspective on behavioral change and new mobility patterns. Such disciplinary perspectives are important, in order to advance knowledge in each of these areas. What is missed, however, is an understanding of the interlinkages between both domains. Whether or not a technological innovation will in fact produce environmental benefits depends to a large degree on the way that it is applied in a concrete social context. In turn, the social context also cannot be characterized as a given, because technological innovations may give rise to changing social practices and preferences (Schwanen et al., 2011, p. 1000). Thus, it is important to gain a better understanding of such co-evolutionary dynamics, especially from a sustainability perspective. There may be mutually beneficial socio-technical dynamics, increasing the potential for better environmental performance, while there may also be detrimental mechanisms where the potential benefits of a technological innovation cannot be realized due to an unexpected way of applying it, or vice versa. These interlinkages are to some extent studied in economic approaches in the field of innovation studies, focusing for instance on dynamics in innovation systems, market dynamics, diffusion and user acceptance. However, these approaches also tend to remain limited to their specific economic context and a market-based logic of supply and demand as the given framework conditions, rather than a part of the research question. Apart from these disciplinary traditions, the specific empirical field of transport and mobility also contributes to the described research gap, because due to the fact that the automobile is both a sacrosanct icon of popular culture and an essential requirement of everyday life, strategies to manage its adverse effects have always been tentative and partial (Cohen, 2006, p. 29). Thus, it might be argued that the technological lock-in of car-based transport systems is coupled with a disciplinary lock-in of research approaches of studying this system. Alternative problem-framing and scientific approaches rarely emerge, especially when these are difficult to quantify with commonly used tools (Schwanen et al., 2011, p. 998). A transitions perspective on system innovation allows for such a broadening of perspective, because studying the emergence of new functionalities through socio-technical co-evolution specifically aims at understanding deep-structural change that crosses disciplinary boundaries and questions societal framework conditions that are usually treated as given. Such a perspective is useful with regard to the research question of this thesis: When interested in the potential of e-mobility as a contribution to more sustainable transport systems, it is important to look beyond processes of diffusion of electric vehicles, because it 10

24 can be assumed that the BEV will not smoothly emerge as a substitute for conventional cars (as indicated by its troubled history) and as a pure substitute it will most likely not produce substantial environmental benefits. Therefore, an assessment of whether e-mobility has system-innovative potential is relevant with regard to the success of electric vehicles, which most likely depends on a changing social context, and with regard to questions of sustainability, which also depend on the emergence of new functionalities as a result of sociotechnical co-evolution. A research approach centered on the concept of system innovation thus allows for studying e-mobility as a new interface between mobility and energy systems, as a trigger for changing perspectives on the role of the car and redefining mobility in general. This already broad approach is further complicated, because the phenomenon of interest, the system innovation, does not exist and it will hardly be possible to judge with certainty whether it will exist in the future, and if so, what it would look like in detail. Nonetheless, considering the current momentum in developing e-mobility and the hopes for sustainable solutions connected to it, it is crucial to find a way of assessing at least the potential inherent in these developments and arrive at recommendations for policy and business for fostering e-mobility as a system innovation. Since transitions, or system innovations, are complex long-term processes emerging from contingent and interrelated dynamics, the question is how to assess the potential connected to a specific technology at an early stage of development. With regard to the relevant variables and influencing factors, [r]esearch shows that it is impossible to isolate any single factor, and that accidental causation has a role to play as well (Schot et al., 1994, p. 1061). It is therefore important to opt for exploratory research designs that do not unduly limit the set of possible influencing factors and that account for contingency by being cautious with extrapolating observed trends. At the core of a system innovation are the co-evolutionary dynamics emerging between a technological innovation and its social context and therefore, in order to assess the potential for system innovation is it imperative to pay more attention to the minutiae of the practices in which technologies become enrolled (Schwanen et al., 2011, p. 1000). This calls for an approach that is case-specific, thus tracing how e-mobility is taken up and dealt with by the involved actors in a concrete setting, and carefully considers the contextual factors having an impact on developments and which may change over the course of time. The central theoretical concept used in transition studies is the multi-level perspective (MLP), which provides a useful heuristic for analyzing system innovations in such a comprehensive way. However, with regard to the research question of this thesis, two specific 11

25 weaknesses need to be addressed. First, since the MLP has been developed based on historical case studies it is difficult to apply it to on-going transition processes. Second, the heuristic character of the MLP precludes a focus on micro-level processes of transitions. This is problematic because a central element of a system innovation is the emergence of new functionalities, i.e. paradigmatic change in the structures and practices of a system as a result of socio-technical co-evolution on a micro-level. In order to address these weaknesses, a theoretical framework is developed in this thesis based on the MLP and introducing the concepts of transformative capacity of a new technology (do BEVs trigger social innovations, e.g. new business models, political regulations or use patterns?) and system adaptability (how inert or open to change is the mobility system?) (based on Dolata, 2009). These concepts provide a way of analyzing the basic dynamics of a system innovation on a micro-level and give an indication of its actual potential in an ex ante situation. This type of exploratory research based on heuristic frameworks is a typical approach in qualitative social sciences and the basic premises of this research tradition apply to this thesis as well. The basic aim of qualitative social research is to study social phenomena in such a way that the complexity involved in human action, societies and specific social groups can be grasped as comprehensively as possible. It is argued that in the social sciences, explaining ( Erklären ) laws and causalities, as in the natural sciences, is not possible, because human behavior and social phenomena simply are not subject to universal laws. The focus is instead on understanding ( Verstehen ) the logic of actions, interpretive schemes and social structures (Flick et al., 2009, p. 14). This does not necessarily mean that in qualitative social research it is generally not possible to test theories and formulate precise hypotheses, as is discussed in controversies about whether fundamental differences exist in the methodologies of quantitative and qualitative research approaches in the social sciences. Following Gadenne (2001), it is argued here that based on a critical and open attitude, research design and methods should always be appropriate in light of the specific research question and object of study (p. 18). This is a scientific imperative, independent of the respective scientific or methodological tradition. In cases where the object of study is not well known and when there is little empirical data, formulating precise hypotheses will usually not make much sense (p. 21 ff.). E-mobility is such a case, at least when approached from a system innovation perspective. Empirical evidence exists with regard to various aspects of e- mobility, e.g. it has been shown that, from a purely technological point of view, the electric vehicle is an old innovation, in economics and innovation studies basic mechanisms in the diffusion of innovations in general, market dynamics or factors determining user acceptance 12

26 of new technologies have also been studied to quite some extent. For these disciplinary research (sub-)questions, formulating precise hypotheses and applying quantitative methods may be useful. However, the scope of the research question addressed in this thesis is broader and interdisciplinary in nature, because it focuses on concepts of system innovation and sustainability, which are inextricably linked to such issues as the organization of contemporary societies, the role of transport therein, justice and ethics (Schwanen et al., 2011, p. 1004). Apart from this broad perspective, the concrete empirical focus of the thesis is on processes of socio-technical co-evolution in the field of e-mobility. These are case- and context-specific and the empirical basis is limited, because e-mobility is still a small market niche and mainly observable in pilot and demonstration projects. Therefore, an exploratory approach is useful, which aims at understanding a phenomenon in depth, in order to identify relevant variables in the first place, which may then lead to the formulation of more general hypotheses. Still, in order to utilize and build on general theoretical insights in the field of transition studies (and its rich theoretical backgrounds in other fields) and in order to avoid the theoretical blindness often critical in grounded theory approaches, basic theoretical propositions will be developed that guide the case study analysis. The basic methodology, the case study design and the combination and triangulation of methods applied will be laid out in detail in chapter 4.1. In sum, the research approach of this thesis is exploratory and builds on a heuristic theoretical framework, in order to meet the requirements of understanding as a crucial principle in qualitative social science traditions. The contribution of this thesis is twofold. First, it aims at providing a theoretical contribution to the field of sustainability transitions research and the research gap identified above, i.e. in terms of providing a more systemic and interdisciplinary perspective on e- mobility as a potential solution for environmental problems in the transport sector. Based on a comprehensive literature review, the various strands of literature dealing with e-mobility are systematized. It can be shown that research on e-mobility is largely structured along disciplinary lines and within sustainability transition studies, which provides a more comprehensive perspective on system innovations, there is a need to refine existing theoretical approaches in such a way that they can improve the analysis of potential system innovations at an early stage. A theoretical framework is developed based on the MLP and Dolata s (2009) concepts of transformative capacity of technologies and adaptability are integrated, resulting in an adapted framework. By facilitating an explicit focus on processes of socio-technical co-evolution related to a specific technology, against the background of multi- 13

27 level dynamics of broader structural change in sustainability transitions, this framework can improve ex ante analyses of potential system innovations. A focus on the concrete interrelations between the introduction of electric vehicles and the way they are dealt with by different groups of actors facilitates a better understanding of the patterns that may lead to successful and system-innovative applications of e-mobility. Analyzing these dynamics with a view to structural change, as conceptualized in the MLP, and guided by an explicit sustainability perspective allows for identifying and assessing in detail the drivers and barriers for e-mobility as a sustainable system innovation. By applying this framework, e-mobility is also introduced in a novel way in transition studies. So far, electric vehicles have been discussed as one possible alternative to ICE cars among many technological options, while the specific potential of e-mobility in combination with newly emerging business models and use patterns (e.g. carsharing or intermodal mobility), and as a technology at the interface of transport and energy systems has not been studied extensively, yet. Furthermore, e-mobility is not the typical green technology niche in transition terms, because in many countries its development is supported by the government and large car manufacturers are involved as is the case in Germany as a prominent example in this respect. Second, this thesis provides an early evaluation of current developments in the field of e-mobility in Germany, which is still at an early stage as regards the public funding programs and demonstration projects starting in Due to the focus on e-mobility as a sustainable system innovation, it provides an addition to conventional accompanying research, which is more disciplinary in nature and focusing on specific aspects of e-mobility in separated research projects (e.g. focusing on life-cycle assessments of electric vehicles, user acceptance, energy management or grid integration concepts etc.). It is thus the first analysis from a comprehensive sustainability perspective evaluating current efforts from a more systemic perspective, thus offering recommendations for policy and business on how to increase the potential for e-mobility to evolve as a sustainable system innovation. In short, it will be concluded that while overall system-innovative potential currently is low, this could be improved by adapting the current funding programs in such a way that actors from outside the realm of (auto-)mobility get involved more ( powerful outsiders, e.g. from the fields of public transport, housing, information and communication technology (ICT)), creating framework conditions that foster direct cooperation between actors from the automotive, energy and ICT sectors, establishing networks with strong and independent coordinating agencies at the center, and by designing demonstration projects in more inter- and transdisciplinary ways. 14

28 1.3 Structure of the Thesis In the following, the structure of the thesis will be outlined in detail and Fig. 2 provides a basic overview of the five main chapters and highlights the basic course of investigation. In chapter 2, e-mobility will be introduced as a field of innovation and a strategy for solving environmental problems in the transport sector. It will be shown that addressing sustainability challenges of modern transport systems requires a broad approach, going beyond technological innovation of vehicles or fuels. It is thus argued that a concept of e-mobility is needed that is not limited to a focus on the diffusion of electric vehicles, but also includes a perspective on long-term and fundamental change in infrastructures, institutions, business models, individual travel behavior and mobility patterns. This sets the stage for the central question of this thesis, i.e. whether e-mobility can potentially emerge as a system innovation and be part of a transition to sustainability. As a starting point, some basic definitions are provided in chapter 2.1 and e-mobility is defined as a system of personal mobility, including different kinds of electric vehicles but also the social context in which they are used. Adding a more specific sustainability perspective, visions and concepts of sustainable mobility are introduced in chapter 2.2 and these are related to the concept of e-mobility as a system innovation. It is shown that applying a broad definition of sustainability in the field of transport requires strategies of increasing efficiency of vehicles and fuels, but also strategies for reducing the volume of transport and achieving modal shift. While e-mobility is mostly discussed as a strategy of increasing resource efficiency and reducing emissions, the aim of this thesis is to connect e-mobility to a broader understanding of sustainable mobility (e.g. new business models in the field of e- carsharing, integrating electric vehicles in intermodal mobility chains and in new forms of city and land use planning). In chapter 2.3 it is shown that a broadening perspective on e- mobility does not only go beyond issues of technological innovation, it even transcends the boundaries of the transport sector. In general, e-mobility can be understood as part of a wider energy transition, moving from fossil resources to renewables, in order to satisfy societal needs for heat, electricity and mobility. At a more concrete level, there are some direct interlinkages: the boundaries between the automotive and the energy sector become blurred in a system where cars are powered by electricity. Furthermore, sustainability transitions in the field of renewable energy and e-mobility depend on each other s success, while at the same time each can significantly contribute to the success of the other: The energy sector by 15

29 providing green electricity for electric vehicles and the automotive sector by providing electric vehicles, the batteries of which can be a means of temporarily storing electricity in an energy system based on fluctuating renewable sources. In chapter 2.4, the complexity of the many interlinkages of e-mobility with general aspects of sustainable mobility and with the energy transitions as an important background is integrated in the concept of e-mobility as a sustainable system innovation. The concept of system innovation is introduced and applied to the case of e-mobility, and a sustainability perspective is added. Based on this, three traps are identified that need to be overcome, if e- mobility is to emerge as a sustainable system innovation. These are, first, a quantitative trap, i.e. there needs to be a certain rate of diffusion of electric vehicles; second, a qualitative trap, i.e. new functionalities need to emerge in the way that mobility needs are fulfilled or the way that electric vehicles are integrated in the energy system, including new business models, use patterns, regulation and symbolic meaning related to mobility; third, a sustainability trap, i.e. new system configurations should deliver sustainability benefits, in the case of e-mobility this would be achieved through renewable energy sources for powering vehicles and integrating electric vehicles in carsharing services or intermodal transport systems. These three traps will in the following serve as a guiding concept for the analysis of the empirical case and for reflecting on the results. In chapter 2.5, a literature review is presented, which aims at identifying whether there are already fields of research or theoretical approaches that focus on e-mobility and relate it to broader definitions of sustainability. The three traps are used to guide and systematize the results of the literature review, in order to show what specific aspects (e.g. technological innovation and diffusion, processes of socio-technical co-evolution, sustainability assessment) are dealt with and asking whether approaches exist that can capture the complex phenomenon of e-mobility as a (potential) sustainable system innovation. It can be shown that currently research on e-mobility is largely dominated by disciplinary perspectives on technological and economic issues relevant for the diffusion of electric vehicles. What is lacking is a perspective on the interrelations between technological change in its social context, specifically tailored to the case of e-mobility, and with an explicit perspective on sustainability. In chapter 3, the theoretical background for the thesis is presented and an analytical framework will be developed for analyzing e-mobility as a sustainable system innovation. The thesis builds on theoretical approaches developed in the field of sustainability transitions 16

30 research. This newly emerging field of research explicitly focuses on sustainability-oriented system innovations, or transitions, and a brief overview will be given in chapter 3.1, including basic definitions and concepts, their origins in other fields of research, and the different theoretical approaches for studying transitions (chapter 3.1.1) as well as a discussion of the role of technology and innovation in this context (chapter 3.1.2). The Multi- Level Perspective (MLP) on sustainability transitions as the central theoretical approach in this field will be discussed in detail in chapter 3.1.3, focusing on how it conceptualizes fundamental change based on Giddens structuration theory (chapter ) and providing an overview of the different dynamics and mechanisms involved in transition processes (chapters ). A critical reflection of the MLP regarding its scientific value and shortcomings in general, and with regard to its applicability to the research question of this thesis will be presented in chapter A major challenge is the question of how to deal with transitions in the making, such as the case of e-mobility. An ex ante analysis is needed and this requires a specific focus to be added or emphasized further in the MLP, which serves as a broad heuristic framework. Chapter 3.2 is dedicated to this issue and two, more future-oriented, approaches in transition studies are introduced, i.e. Strategic Niche Management (chapter 3.2.1) and Transition Management (chapter 3.2.2). It will be shown that these approaches focus on processes of steering and managing transitions in practice (chapter 3.2.3), while they are less suited as analytical tools assessing on-going developments. The aim of this thesis is to find a way of analyzing the potential inherent in currently observable dynamics in the light of an envisaged sustainable system innovation (chapter 3.2.4). Therefore, a theoretical framework building on the MLP and addressing its weaknesses discussed earlier is developed in chapter 3.3. Based on a discussion of the relevant unit of analysis in studying system innovations (chapter 3.3.1) and a specific conceptualization of system innovation (chapter 3.3.2), Dolata s (2009, 2011, 2013) concepts of the transformative capacity of new technologies and adaptability are introduced in chapters and An integrated analytical framework will be developed in chapter by integrating these concepts into an overarching multi-level perspective on transitions and it is shown that this provides a way of analyzing the basic dynamics of a system innovation at a micro-level and give an indication of its actual potential in an ex ante situation. 17

31 In chapter 4, the integrated analytical framework will be explored by tracing the dynamics of the German innovation system for e-mobility, focusing especially on the automotive state of Baden-Württemberg. The aim is to gain insights regarding the conditions for increasing system adaptability or barriers impeding it as a way of assessing the potential of this particular e-mobility niche and its chances for future up-scaling. The methodology guiding the empirical analysis is laid out in chapter 4.1. The analysis proceeds along three steps. First, the theoretical concept of a sustainable system innovation will be delineated as an ideal-type for the case of e-mobility in chapter 4.2. To that end basic criteria of sustainable forms of e-mobility will be deducted from the discussion of sustainable mobility in chapter 2 and these criteria will serve as an orientation for the overall analysis and as selection criteria for system-innovative e-mobility projects later in the case study. Idealtypical elements of sustainable e-mobility identified in that way are the emerging interlinkages between e-mobility and the broader context of a transition to renewable energies (chapter 4.2.1), e-mobility and less car-dependent forms of mobility in fleet applications (chapter 4.2.2), intermodal transport concepts (chapter 4.2.3), and innovative business models in the field of carsharing (chapter 4.2.4). Second, the transformative capacity of the battery-electric vehicle will be assessed in chapter 4.3. Based on a set of qualitative criteria developed in the theoretical framework it will be analyzed in what ways the BEV affects the technological profile, institutional structure, patterns in market relations, R&D, policy making and user behavior, patterns of interaction, and the boundaries of the socio-technical system of (auto-)mobility. Since the electric vehicle is an old innovation, secondary sources exist on these issues and are used as the basis of the analysis. Third, the concept of system adaptability will be explored in a case study of e-mobility development in Germany, and Baden-Württemberg more specifically in chapter 4.4. The basic rationale for choosing a case study has been laid out in the methodology section and the concrete case study design, the combination and triangulation of methods is described in chapter An overview of the German innovation system for e-mobility is presented in chapter with a particular focus on political initiatives at the federal level (chapter ), the actor structure determining the strategic perspective of the German government (chapter ), and the development of funding programs (chapter ). A closer look will then be taken at the automotive state of Baden-Württemberg in chapter A brief overview is given on the state of Baden-Württemberg in general and the e-mobility initiatives taking place during the period of investigation from 2009 until today (chapter ). In a 18

32 network analysis, the dominant actors and cooperation patterns in the context of developing e- mobility in Baden-Württemberg will be analyzed (chapter ). Subsequently, basic frames of reference in relation to e-mobility shaping the overall discourse in this network are analyzed, in order to complement the insights on different levels of influence of specific actors with their respective views, guiding principles and the resulting strategies (chapter ). In chapter , a closer look is taken at specific e-mobility projects carried out in Baden-Württemberg that can be characterized as system-innovative. These projects are selected based on the ideal-typical criteria identified for e-mobility as a sustainable system innovation and analyzed in order to show where and how concrete system-innovative approaches emerge (e.g. regarding involved actors, cooperation patterns, motivations and goals). In-depth expert interviews have been carried out with actors involved in these projects and with actors responsible for coordinating the innovation system as a whole. The results of the case study of Baden-Württemberg are presented in chapter and they are discussed in the larger German context of developing e-mobility in chapter The empirical findings are then discussed in relation to the theoretical framework and reflected with regard to the three traps of a sustainable system innovation in chapter 4.5. The results of the case study are taken up and discussed against the background of the theoretical concepts of transformative capacity and system adaptability and the basic theoretical propositions regarding the potential for system innovations to emerge based on the MLP. The case of Germany provides a number of lessons as regards the potential of e-mobility to develop as a sustainable system innovation and these will be discussed here. For instance, the analysis has shown that the system-innovative potential of this specific e-mobility niche remains limited, due to the powerful influence of incumbents, conflicting political goals and traditional science approaches. A few more system-innovative activities emerge where powerful actors from outside are involved, who are capable of viewing mobility in a more systemic way (e.g. actors from the public transport, ICT or housing sector). It is argued that the role of large demonstration projects is important, but they need to be designed as transdisciplinary research projects from the beginning. Conclusions will be drawn in chapter 5 and an outlook is presented, including recommendations for improving R&D funding programs and demonstration projects for developing e-mobility as a sustainable system innovation as well as future research needs in the field of transitions to sustainable mobility. 19

33 Fig. 2: Structure of the Thesis - Main chapters and course of investigation. 2 E-MOBILITY AS A SUSTAINABLE SYSTEM INNOVATION Mobility plays a special role in the debate about future sustainable development and the great transformation of modern societies. This is not only due to the fact that the transport sector has a large direct and negative impact on global eco-systems, but also because mobility is such an integral part of modern societies and is thus an element of the root causes and hazardous effects of current unsustainable development. As Høyer puts it, the fossil society and the mobile society have grown like Siamese twins (2008, p. 68). Modern transport systems facilitate a globalized economic system, the car industry and its related sectors and supplier networks are of key importance to many national economies, urban infrastructures have facilitated as well as themselves been shaped by the increasing dominance of car-based individual mobility, and the car as an artifact is a symbol of modernity and prosperity (Cohen, 2006, p. 29; Norton, 2011; Pel & Boons, 2010, p. 1250; Rammler, 2011, p. 23). These many interlinkages of modern forms of mobility with various aspects characterizing (mostly Western) industrialized economies make it extremely difficult to envision substantial change and shows that a broad perspective is needed. At the same time, this also means that a 20

34 transformation towards a more sustainable form of mobility may serve as a Trojan horse for an overall transformation towards a more sustainable, decarbonized society. Especially an electrification of transport systems should then be advanced in terms of a change from a fossil to an overall solar culture (Weider & Rammler, 2011, p. 6). The need to reduce the environmental burden caused by currently fossil-based transport has been recognized and the pressure exerted by increasingly strict environmental regulation, e.g. regarding CO 2 -emissions, has led to at least some concern and efforts towards sustainability-oriented change. Typically, this type of change is hoped to be achieved with the help of technological innovations such as the battery-electric vehicle (Høyer, 2008, p. 68). However, in most industries, technological innovations usually follow so-called technological trajectories. They build on the existing technologies in their design and function and often remain limited to incremental changes in this basic technology, an add-on or particular improvement. Thus, innovations in this sense seldom have a substantial impact on the already existing infrastructures, use patterns or business models related to a specific technological field (Elzen & Wieczorek, 2005, p. 653). As in many other cases, such a technological trajectory can be observed in the automotive industry. The basic design of the car and its core technology, the internal combustion engine, has essentially remained the same (Weisshaupt, 2006, p. 22). Numerous innovations have been developed, that can be characterized as linear, in the sense that they have improved the basic technology in various ways. Engines have become more powerful and more efficient, safety has been improved through airbags and electronic driving assistance systems, environmental performance has been improved by introducing the catalytic converter and exhaust filters (Weisshaupt, 2006, p. 85 f.). At present, many of these innovations are aimed at meeting increasingly strict CO 2 - emission standards set by the EU as well as national governments and, more generally, find a way of making motorized individual transport more sustainable. However, the effects of technological improvements have begun to reach their limits, due to the absolute number of vehicles and volume of traffic and transport, which continue to grow larger. As a consequence, the total environmental impact of cars is increasing in absolute numbers. Apart from that, air pollution and emissions are not the only issues that have to be tackled. Due to its oil-dependency, the modern car-based transport system will in the long run be affected by resource scarcity. From a global point of view, it seems doubtful whether broad access to mobility in large and livable cities can be possible at all with a car-based system of individual mobility, which involves increasing levels of congestion, of local air pollution and an 21

35 increasing need for space taken up by parked cars (Elzen & Wieczorek, 2005, p. 653 f.; Köhler et al., 2009, p. 2986; Weisshaupt, 2006, p. 86). Battery-electric vehicles could solve part of these problems, especially those related to local air pollution, CO 2 -emissions (depending on the sources of the used electricity) and oildependency, since in this particular respect they provide an alternative to conventional cars powered by fossil fuels. A number of problems impede their widespread introduction and diffusion, though. First, from a consumers point of view, BEVs do not perform as well as conventional cars, especially with regard to vehicle range and charging time. Second, the charging infrastructure for BEVs needs to be built up in the first place. Third, BEVs involve high costs for consumers and producers since in this early phase they cannot benefit from economies of scale (van Bree et al., 2010, p. 534). These problems are generally viewed as the typical challenges of an immature technology. However, even if these could be solved, by improving battery technology and developing fast-charging systems, the mere diffusion of BEVs does not guarantee a sustainability benefit per se. It may even worsen the situation in the transport sector by adding vehicles that are energy- and CO 2 -intensive, depending on the electricity source for powering BEVs (Schallaböck et al., 2012; Zimmer et al., 2011). Another problem is resource intensity, especially when considering the production of batteries (Kushnir & Sanden, 2012, p. 102; Stamp et al., 2012, p. 104) and the potential of BEVs being used as the eco-friendly additional private car. Finally, issues such as congestion and sustainable mobility in dense urban areas would remain untackled (Vallée & Schnettler, 2013, p. 34 f.). This implies that a more radical change is needed: The focus on a technological innovation, such as the BEV (or, for that matter, biogenic fuels, hydrogen fuel cell vehicles etc.), will hardly suffice to deal with the problems inherently connected to the modern carbased transport system. What is needed is a system innovation, which can be defined as a combination of technological and social innovations that merge into a co-evolutionary process of substantial change in a socio-technical system (such as the car-based transport system) over the long term (Geels, 2005a, p. 6, p. 10 ff.). In the case of transport, a system innovation would include new vehicle technologies, new infrastructures, new logistics, new legislation and change of travel behaviour (Elzen & Wieczorek, 2005, p. 654). In this sense, the buzzword e-mobility already points to a system innovation that transcends the electrification of the car and its technological implications (Canzler & Knie, 2011, p. 101 ff.; Rammler & Sauter-Servaes, 2013, p. 48). The question thus is, can e-mobility be more than 22

36 the introduction of BEVs and rather a system innovation towards a future, more sustainable mobility system? 2.1 Definitions The aim of this section is to clarify what is meant by electric mobility or e-mobility in the context of this thesis. The common understanding of e-mobility is in terms of a shorthand for electrified automobility. The technological artifact at the center of this type of mobility is a vehicle that is driven by a battery-powered electric motor, the battery-electric vehicle. The term usually includes all road vehicles with at least three wheels, thus including different kinds of private cars, commercial vehicles and busses (Sauter-Servaes, 2011, p. 25; Weider & Rammler, 2011, p. 5). The most significant difference of the BEV as compared to other types of alternative drive technologies or fuels is that it can be characterized as a radical innovation that differs from conventional vehicles with regard to basic characteristics and ways of functioning. The current system of personal mobility is built around privately owned cars, which are designed as general purpose vehicles, i.e. fulfilling all kinds of mobility needs in terms of length and purpose of trips. They are embedded in an extensive road, parking and refueling infrastructure that facilitates established patterns of motorized private transport. The basic characteristics of this system are not in principle challenged by the introduction of biofuels, for instance. Even though in this case current fuels would be replaced and new fuel producers would enter the market, the basic infrastructure and use patterns would remain intact. With regard to alternative vehicles, the fuel cell vehicle (FCV) challenges the system in terms of the established drive technology and it would require a completely new refueling infrastructure based on hydrogen filling stations. However, FCVs would not amount to more than a replacement of one sort of general-purpose vehicle by another. In principle, the way that FCVs could be used and marketed would not differ from today s cars. They can be used for short and long trips and the refueling infrastructure and process would function much like today s gas stations. Similarly, the different types of hybrid electric vehicles are also all aimed at fulfilling the same needs as a conventional car powered by an internal combustion engine. This is not to say that the diffusion of FCVs or hybrid electric vehicles is not a challenge. It would require various technological innovations in drive technology and new forms of cooperation as well as an adjustment of industry structures to create the links with electricity or hydrogen as power sources. Nonetheless, the basic infrastructure design, mobility patterns 23

37 on the part of users and business models on the part of car manufacturers would remain similar (Nilsson et al., 2012, p. 61; Weider & Rammler, 2011, p. 6). In contrast, the pure BEV is much more difficult to integrate into the current transport system. It will most likely not become feasible for a BEV to compete with a conventional car powered by an internal combustion engine with regard to range and price, thus with the currently dominant idea of the car as a universal, all-purpose vehicle. The BEV thus challenges the dominant paradigm where personal mobility is guaranteed by the privately owned car. While any type of a hybrid or fuel cell electric vehicle rather represents an evolutionary or even incremental innovation, in the sense that they fulfill the same functions as a conventional car in similar ways, the pure BEV points towards a more paradigmatic change in the transport system (Weider & Rammler, 2011, p. 6). This is not to say that the BEV, once it emerges, will automatically lead to fundamental changes in modern transport systems. From a historical point of view it can be shown that the BEV has failed before. However, this also shows that the BEV can only be successful when evolving in terms of a more systemic change towards e-mobility. The technological shortcomings of the BEV as compared to conventional cars in the context of today s car-centered transport system become irrelevant or may even turn into advantages when the BEV is envisioned as part of a system innovation towards e-mobility (Weider & Rammler, 2011, p. 6, 10). The term e-mobility is thus used to capture the concept of such a paradigmatic change, or the vision of it, which goes beyond a technological innovation, in the sense that it has implications for the way this innovation is used, marketed, regulated and embedded in cultural contexts. Therefore, the concept of e-mobility used in this paper focuses on the BEV as the central technological artifact around which a certain type of mobility evolves. This may potentially take forms that are markedly different from current mobility patterns dominating the transport system. The BEV may contribute to a transition towards new forms of mobility, because it can be characterized as a radical innovation. The term e-mobility consequently captures a multi-facetted network of innovations around the car as a product, the way it is used and embedded in spatial, social and cultural structures (Rammler, 2011, p. 14; Schneidewind, 2011; Weider et al., 2011a, p. 52). Such a perspective allows a perspective on the BEV where it can be viewed as an artifact that transcends traditional sector boundaries and connects a society s need for energy and mobility. Visions of new forms of urban mobility and innovative mobility services embedded in new energy infrastructures emerge around the BEV (cf. Dennis & Urry, 2009; Mitchell et al., 2010). This also implies that e- 24

38 mobility as a system innovation takes the BEV as point of departure but also needs to include new types of vehicles, such as pedelecs or segways, as well as long-established forms of electrified mobility, such as trains and trams (Weider & Rammler, 2011, p. 6, p. 10). For the purpose of this thesis, e-mobility is defined as a system of personal mobility where different types of battery-electric vehicles have replaced conventional cars and have become part of an overall electrified transport system, and where electric vehicles have become embedded in new transport infrastructures and policies, behavioral patterns by users, market and industry structures as well as mobility cultures. In order to focus on personal e- mobility, freight transport is excluded and the focus on battery-electric vehicles is broadened insofar as public transport or other electric vehicles, such as trains or pedelecs, are considered as parts of the mobility system in general and possibly as elements that gain importance in a future sustainable mobility system, where the role of the privately-owned (electric) car loses importance. 2.2 Visions and Concepts of Sustainable Mobility Apart from a systemic perspective on e-mobility, including the infrastructural, socioeconomic and cultural context of the battery-electric vehicle as a technological innovation, a view on sustainability needs to be included, if new forms of e-mobility are to contribute to the solution of environmental problems in the field of transport. Therefore, the concept of e- mobility as a system innovation needs to be connected with concepts of sustainable mobility in general. Awareness for environmental problems related to transport rose significantly from the 1960s onwards, especially in Western countries. In the beginning, negative effects of transport were perceived largely in terms of pollution in local settings and as a problem of the intensity of transport, especially in cities. With increasingly growing volumes of transport, problem perceptions broadened and apart from local pollution also included emissions by the transport sector on a global level, continued increases in energy demand for passenger and freight transport as well as problems of congestion (Høyer, 2008, p. 68). The environmental impact of transport on global ecosystems is severe: The amount of energy consumed in the field of transport amounts to almost 30% of total global energy consumption (IEA, 2013, p. 246). Fossil fuels have remained the dominant energy source, with approximately 95% of energy use in the field of road transport being based on oil - a share similar to the situation in the 1970s (p. 510). Thus today, the transport sector, together with the generation of electricity and heat, are responsible for more than 50% of the EU s total CO 2 -emissions (EEA, 2013, p. 25

39 14 f.). Looking at the car as the dominant means of individual transport, it can be shown that from a life-cycle perspective on an average car, not only emissions and pollution caused during its use are critical. In the production stage of one average car 26.6 tons of waste and 922 m 3 of polluted air have been produced, and, this production phase preceding actual car use, together with the final disposal stage, cause 20% of a car s total environmental impact (Orsato & Wells, 2007, p. 1001). Apart from such direct environmental harm induced by modern transport systems, the problem of congestion, especially in cities, also amounts to a severe hazard caused by modern transport systems. The International Energy Agency projects that between 2011 and 2035, the number of vehicles worldwide will have doubled (IEA, 2013, p. 71). In accordance with this broadening perspective on the many dimensions of problems that modern transport systems are facing, envisaged strategies for improvement have also evolved, from technological innovation and improving ecological performance in industry supply chains (cf. Seuring et al., 2003) to a broader outlook on mobility per se. While in the beginning the focus was predominantly on technological improvements of vehicles, fuels and emission control, increasingly broader perspectives have begun to also look at the infrastructures, politics and lifestyles that are connected and intertwined with the transport sector causing not only environmental problems, but also social problems regarding access to mobility, land use and livable cities (Cohen, 2006, p. 27). Thus, as a result of the multitude of problems connected to modern transport systems, broader visions of sustainable transport or sustainable mobility emerged in public and academic discourses. Even though sustainability can mean a plethora of different things, it is commonly understood that sustainable mobility should provide social and economic welfare, in terms of societal access, personal mobility and trade, without damaging the environment, in terms of emissions and pollution as well as resource depletion, or otherwise causing harm to humans (Nykvist & Whitmarsh, 2008, p. 1373). While the term sustainability per se includes such a holistic perspective, it makes a difference whether it is applied in terms of sustainable transport or sustainable mobility. While the focus on sustainable transport mainly relates to specific types of vehicles and transportation infrastructures, sustainable mobility is a more comprehensive concept with a focus on the societal context within which people or goods are moving (Høyer, 2008, p. 68; Schneidewind & Zahrnt, 2013, p. 88 f.). In the seminal paper by Banister (2008), sustainable mobility is introduced as a concept that amounts to a new paradigm in the theory and practice of transport policy and planning. According to this approach, strategies for increasing the sustainability of modern transport systems should be adopted across four basic principles or action domains. First, the 26

40 principle of substitution is the basis for strategies aiming at a reduction of the need to travel. In a situation where the sheer volume of transport is problematic, it seems obvious and straightforward to directly decrease this high volume by decreasing the amount of trips undertaken. Developments in the field of Information and Communication Technologies (ICT) are important in this context because they facilitate virtual forms of communication across spatial distances, thus, in principle, reducing the need to travel in order to communicate face-to-face. Substitution therefore means that a specific trip (to work, for shopping or leisure activities) is replaced by a technology or an activity that does not involve the need to travel (Banister, 2008, p. 75). However, some caution is warranted, since the possibilities offered by various ICT applications may also generate new travel activity or lead to a substitution of many shorter trips by fewer, but longer trips (Lyons & Kenyon, 2003, p. 30; cf. also: Banister & Stead, 2004; Kenyon & Lyons, 2007). Second, where trips cannot be replaced entirely, sustainability-oriented strategies should aim at encouraging modal shift, i.e. a shift to other means of transport that have less detrimental effects on the environment. The basic strategy thus is to make the trips that cannot be replaced entirely as environmentally friendly as possible. In our current mobility system this basically means a reduction in the use of cars and a shift towards public transport or even cycling and walking (Banister, 2008, p. 75; cf. Böhler-Badeker et al., 2012; Bongardt et al., 2013; Koska & Böhler, 2008; Reutter, 2010; Reutter & Jansen, 2008; Wilke & Böhler, 2007). Third, strategies addressing the infrastructural context in which mobility takes place, i.e. city and land use planning, should aim at reducing trip lengths. This can be achieved by designing infrastructures and buildings in such a way that density and concentration is relatively high and that different functions (work, home, leisure) can be fulfilled within mixed-use areas. Such planning strategies are aimed at creating high-quality environments especially in cities and they are linked with strategies of substitution and modal shift. Intelligent planning of urban infrastructures would mean close spatial proximity of various day-to-day activities, thus making some trips superfluous and facilitating walking and cycling as well as an increased use of public transport as a more attractive option as compared to the currently dominant use of cars (Banister, 2008, p. 73 f.). Finally, a comparatively moderate and less radical approach for achieving more sustainable forms of mobility is an efficiency increase in the transport system. This entails the introduction of cleaner technologies in vehicle design, drive concepts, fuels and renewable energy sources for powering vehicles. Such technology-based strategies can be combined 27

41 with efforts in city and land use planning, for instance, by restricting access to certain areas for vehicles that are seen as relatively less environmentally-friendly (Banister, 2008, p. 75). Since the focus of this thesis is on e-mobility as a new form of mobility evolving around electric vehicles and the central question relates to the potential that this development entails for the emergence of future more sustainable mobility, issues and strategies of city and land use planning and substitution of trips through ICT applications fall outside the scope of this thesis. They are important for sustainable mobility in general, but for the case of e- mobility they are dealt with in a more implicit fashion, in terms of framework conditions for concrete mobility patterns evolving around electric vehicles. The other strategies of increasing overall efficiency through the use of alternative technologies and energy sources and encouraging modal shift are at the center of this thesis and their interlinkages, i.e. between technological innovation and changes in mobility patterns, are of particular interest here: Instead of using the car for the majority of trips, this would mean that people use combinations of different transport means to organize their specific trips in the most efficient ways. Apart from changes in infrastructure and technology (Lechtenböhmer et al., 2010, p. 7 ff.), this would especially require information services assisting in the organization of individual trips and changes in individual social practices with regard to travel behavior. This reduces the total negative impact of the most disadvantageous means of transport (with regard to emissions and energy usage) and increases overall efficiency of transport, when there is substitution of e.g. car use by a combination of it with walking, cycling and using public transport. In combination with increased forms of collective car ownership or carsharing services, an increase in intermodal mobility patterns would further reduce the resourceintensity of transport and would also further contribute to solving problems of congestion in urban areas. Considering the multidimensionality and severe nature of sustainability challenges the transport sector is facing, it is widely accepted that a purely technological solution to these problems, e.g. through alternative fuels or drive technologies, is not feasible. While fundamental changes in urban infrastructures can only develop over the long term and a substitution of transport via ICT applications would be part of larger societal trends and developments, encouraging modal shift is perceived to be an approach towards overall sustainable mobility that seems more realistic and achievable in a short- to mid-term perspective. Strategies for achieving a shift to less harmful modes of mobility, e.g. walking, 28

42 cycling and public transport, are usually combined with concepts of mobility services 1. These include carsharing, thus integrating cars in multi-modal mobility chains, or software based services assisting in planning and paying for trips using different means of transportation in combination. A central role is attributed to this type of integrated intermodal mobility and here infrastructural and ICT-related aspects do play a role again. In a system where people rely on various means of transportation to make a specific trip ( chain mobility ), there is a need for facilitating suitable infrastructures, e.g. convenient stations for switching between modes, special parking spaces for carsharing systems, bus or car pool lanes, and ICT solutions, such as integrated internet platforms or smartphone applications for planning trips, making reservations, paying for trips and using information services (Cohen, 2006, p. 33; Kemp & Rotmans, 2005, p. 34; Vergragt & Brown, 2007, p. 1105). An important aspect with regard to these types of strategies towards more sustainable mobility has not been mentioned explicitly so far. How can they be implemented in practice? The difficulty in changing the transport sector is not only a challenge in terms of the multitude of problematic aspects and their interlinkages, but also in terms of addressing these various problems in a coherent and comprehensive way (Cohen, 2006, p. 33; Leerkamp, 2014, p. 85 f.). All of the strategies outlined above, except maybe for increasing overall efficiency, are questioning fundamental societal values (e.g. decreasing the overall volume of travel is in conflict with a globalized economy and a global society characterized by individualization, especially in the Western world), they require fundamental rethinking of the way that cities and transport infrastructures have been planned and constructed over decades, and they are in conflict with very personal emotions connected to the car (e.g. the car as a symbol of personal freedom, status and wealth). What is more, even a strategy of increasing efficiency by introducing alternative vehicles or fuels constitutes a tremendous challenge maybe not in all cases connected to fundamental guiding principles of modern societies, but at least requiring severe restructuring of industries and markets (Berg & Schneidewind, 2013). These paradigmatic challenges involved in making the transport sector more sustainable are reflected in the way that the literature deals with sustainable mobility. There seems to be general agreement on the nature of the problems involved and on the ways that these should be addressed in terms of overall strategy and with regard to the range of policy instruments that can be chosen from. However, the empirical reality simultaneously shows 1 Cf. Wilts & Rademacher (2014), Wilts et al. (2013), Leismann et al. (2012) for a discussion of the potential of increasing resource-efficiency and preventing waste through product-service-systems. 29

43 that not much is changing in practice and that there are many well-studied barriers for overcoming the current transport system centered on personal mobility guaranteed by the privately owned, fossil-fuelled car. Banister suggests that reducing car use cannot be achieved directly, but that there need to be more indirect push and pull measures that make other options more attractive and that, in general, people should be involved much more in research and policy making for sustainable mobility (Banister, 2008, p. 75 f.). In sum, a transition towards sustainable mobility would include a number of aspects. First of all, the overall volume of transport would have to be reduced. This means that the need to travel should be reduced where possible (for instance by substituting travel needs with the help of the internet and ICT) and average trip lengths should be reduced, which depends largely on adequate forms of city and land use planning. Furthermore, efficiency of engines and fuels should be increased, in order to decrease the negative impacts of the amount of transport that cannot be reduced or substituted as much as possible. Finally, modal shift should be encouraged that reduces the level of car use and increases the proportion of walking, cycling and using public transport in the overall volume of traffic. The definition of sustainable mobility used in this thesis and the envisaged solutions or strategies for achieving it are specific insofar as they can be grouped among broad conceptualizations of sustainability in contrast to narrow sustainability definitions focusing on environmental impact of specific technologies in terms of emissions and ecological footprint. Research based on such a narrow conceptualization is typically more technology-oriented, rather employing quantitative and scenario analysis. While this type of research is no doubt important, it tends to disregard social aspects and misses the many interlinkages between technologies and their social and institutional surroundings (cf. Wilke, 2013). It may thus be argued that an exclusive focus on alternative fuels and vehicle technology addresses issues of sustainable mobility in only the shallowest and most delimited sense (Cohen, 2006, p. 34). Apart from that, a central argument in this thesis is that achieving even a narrowly defined form of sustainable e-mobility will not be feasible with the help of narrow approaches, i.e. narrowly delineated research questions and approaches staying within disciplinary academic boundaries. This is simply due to the fact that the environmental impact of modern transport systems is caused by the sheer volume of transport and the interconnectedness of technological issues related to vehicles and fuels with issues of lifestyles and social practices in modern societies, the necessities of everyday-life as much as 30

44 political and economic institutions. Thus, aiming at reducing the ecological footprint of transport systems will necessarily require a socio-technical, broader approach towards sustainability aspirations. Such a broader perspective can nonetheless mean that the starting point for analysis is the car as the central technological artifact in modern transport systems and especially personal mobility patterns. Technological lock-in and path dependencies have historically evolved with the automobile, extending over various dimensions ranging from infrastructures, vehicle design heuristics and engineering principles to markets and business models, policy, land use patterns, city planning and modern lifestyles (Cohen, 2006, p. 29; Pel & Boons, 2010, p. 1250). One may argue that in debates about sustainable mobility, the car is the boundary object that connects all these elements and to which actors from different fields can relate in their specific ways. In this context, it is important to note that the car is not only a technological, but also a cultural artifact, and that it is therefore important to address also the symbolic functions connected to cars, e.g. as status symbols or otherwise emotionally important objects (Barth & Jonas, 2011, p. 76). The car and the historical development of an automobile society can thus also be an important starting point for discovering the root causes of the complex and deeply entrenched causes of current unsustainability. This may in turn mean that strategies towards more sustainable mobility systems require a move away from car-centered forms of mobility which is implied by the principles of sustainable mobility discussed above. If car-dependency is indeed at the heart of the problem, then any technological optimization strategy or problem-solving approach working around the car, without at least questioning its role in modern transport systems and societies at large, may remain limited to a self-defeating pseudo-solution (Pel & Boons, 2010, p. 1254). In sum, the general problem underlying the challenges as well as the advocated solutions for achieving sustainable mobility, is the need for paradigmatic change overcoming current forms of car-centered personal mobility and leading the way towards a post-fossil society. 2.3 E-mobility and the Energy Transition Placing e-mobility in the context of larger transition processes and visions of a post-fossil society further broadens the perspective of this thesis: e-mobility is not only more than a project of technological innovation and diffusion, it also reaches beyond the boundaries of the transport sector. Especially in political discourses in the context of the German energy transition ( Energiewende, i.e. the political project of restructuring the German energy system towards a larger share of renewable energies and increased energy efficiency), it is 31

45 often argued that an integrated approach is needed, focusing on e-mobility and renewable energies as natural partners (BEE, 2010, p. 3). The basic challenge is the same for the energy and the transport sector: both are faced with pressure for fundamental restructuring, if they are to contribute to future sustainable development. For both sectors this means not only technological change, but also a redefinition of basic guiding principles, new business models, use patterns, regulation and socio-cultural embedding (Schneidewind et al., 2013). Apart from those common challenges, the envisaged energy and mobility transitions are mutually interlinked at a more concrete level where the boundaries between the automotive and the energy sector become blurred in the actual provision of electricity for electric vehicles. Fig. 3: Mutual dependency and interlinkages in the field of energy and mobility transitions. The BEV is the tangible artifact at the interface of the energy and the transport sector and it illustrates and embodies not only the interlinkage between the two, but also their mutual interdependence: From a transport perspective, the introduction of electric vehicles would increase overall vehicle efficiency and contribute to a reduction of local pollution and noise. However, a more significant and meaningful contribution to a sustainable transport system would require that electric vehicles are powered by electricity from renewable sources thus, 32

46 e-mobility for sustainable development depends on a successful energy transition. In turn, the success of a wider energy transition crucially depends on the availability of short-term storing capacity for electricity from fluctuating renewable sources, which could be provided by the batteries of electric vehicles. Also, the term energy transition already indicates that this is an endeavor of restructuring the way societies produce and consume energy in general, not limited to the production of electricity, and definitely including the way we consume energy in the transport sector. Thus, in short, the energy and the mobility transition depend on each other s success, while at the same time each can significantly contribute to the success of the other (Rammler & Sauter-Servaes, 2013, p. 48). In practice however, the case of Germany shows that energy transition is often dealt with in terms of an electricity transition. It is somewhat neglected that another major field where primary (fossil) energy sources are being used and lead to greenhouse gas (GHG) emissions is the transport sector. In Germany, roughly 160 million tons of CO 2 are emitted by the transport sector each year, which amounts to approximately 18% of Germany s total GHG emissions (UBA, 2012, p. 44; UBA, 2013). Crude oil is the basis of conventional fuels, such as gasoline, diesel or kerosene, and is the primary energy source (90%) in the transport sector (Vallée et al., 2013, p. 77). Thus, a comprehensive and meaningful energy transition requires including a mobility transition. What would this imply in practice? The basic idea is that introducing e-mobility based on renewable energy sources is a way of making transport more sustainable, while at the same time this could in turn contribute to an overall energy transition. The major aim of reducing CO 2 -emissions caused by transport requires that BEVs are powered by electricity produced from renewable sources, such as solar or wind. At the same time, BEVs could become an important element fostering a transition towards a more decentralized energy system based on renewable sources. Their batteries could be used as small-scale storage units for electricity, in order to deal with the fluctuations in solar or wind powered electricity generation. In principle, BEVs could flexibly be charged when there is excess production of electricity and low demand, while they could feed electricity back into the grid when there is excess demand. Currently, the amount of BEVs is still too little and the feasibility of such vehicle-to-grid (V2G) concepts depends on the development of smart grids and smart metering as well as billing systems (Dijk et al., 2013, p. 142; Rammler, 2011, p. 18 f.). Nonetheless, there is a great potential in this field and it shows that there are many concrete interlinkages between sustainability transitions in the field of transport and energy. 33

47 Whether this potential can successfully be exploited largely depends on a systemic approach for the integration of electric vehicles in the energy system. If electric vehicles are used as a means of providing grid stability, there needs to be a mechanism by which it can be guaranteed that electric vehicles are preferably being recharged when there is a high volume of electricity feed-in by photovoltaic systems or wind power stations. Various approaches are being discussed for managing such controlled charging as a contribution to stabilizing energy infrastructures needed in a situation where the share of renewable energy increases. Integrating electric vehicles in this way, i.e. as a buffer for fluctuating electricity feed-in, is already feasible with existing technologies and would already have meaningful effects with a relatively limited amount of electric vehicles, such as, for instance, the 1 million electric vehicles that the German government aims to have on German roads by The more advanced form of integration would be in the form of V2G concepts. Using electric vehicles not only as buffers taking up excess electricity but also as bi-directional storage elements that feed electricity back into the grid when needed, is much more technologically demanding. It also requires more advanced forms of controlled charging. This can be realized through flexible electricity tariffs that are adjusted to the respective grid situation and volume of electricity generation. An option may also be to include electric vehicles as power reserve in existing policy instruments fostering renewable energies, such as the German feed-in tariff by, for instance, providing compensation for guaranteeing grid stability. These measures show that it is central to adopt a systemic perspective on the electric vehicle and its role in the context of the energy system, in order to realize the potential of an integrated mobility and energy transition (Canzler & Knie, 2011, p. 48 f., p. 51; Turton & Moura, 2008, p. 1106; Vallée & Schnettler, 2013, p. 35). A number of problems related to such controlled charging schemes point to the fundamental challenges that a more systemic approach is faced with and that highlight that the seemingly obvious synergies in achieving a mobility and an energy transition are impeded by deeply embedded structural constraints. For instance, the expectation that energy utilities would invest in public charging stations for electric vehicles has not been met so far. The current business model of these companies would require that this infrastructure investment could pay off via electricity contracts with owners of electric vehicles. This will not be feasible, not only because of the currently low numbers of electric vehicles in the first place, but also because refinancing expensive public charging stations would require electricity prices significantly exceeding household prices, thus the price for charging electric vehicles at home (Canzler & Knie, 2011, p. 66 f.). On the part of car manufacturers, the feasibility of an 34

48 integration of BEVs in the energy system would require the technological development of batteries and battery management systems in vehicles to focus on the ability to function as a reliable means of storing electricity and feeding it back into the grid. However, the focus of R&D activities is much more on increasing range, longevity and cycle stability of batteries, because the business model of car manufacturers requires them to develop electric vehicles that can compete with conventional cars (Canzler & Knie, 2011, p. 54). Similarly, if electric vehicles are used in just the same way as today s conventional cars, it is doubtful whether private car owners would be willing to accept the restrictions that go along with controlled charging, i.e. charging preferably at certain times and making the car available as an electricity source at others (Canzler & Knie, 2011, p. 49 f.). These very practical challenges clearly show two things. First, the energy transition and the envisaged goals of electrified mobility are interlinked, they cause hope for mutual synergies and benefits, and they even depend on each other for a more widespread success in both arenas. Second, however, it also becomes obvious that the potential that lies in these interlinkages requires a fundamental rethinking of various aspects that tend to be taken for granted. The way that an energy and mobility transition could materialize seems realistic only if there is room for adapting very basic assumptions in the way that business models for car manufacturers and energy utilities are being developed, the kinds of policy instruments that would be needed to foster systemic integration and, last but not least, the way that the car is perceived and used everyday. It would in fact amount to a fundamental transformation of the firmly locked-in fossil fuel-based modern energy systems with their long-established and firm ties between fossil energy sources, the way they are produced and marketed, the way they provide the foundation for various societal subsystems (not just the energy and transport sector), and the resulting lifestyles (Meadowcroft, 2009, p. 329). The severity of this challenge also implies that the electrification of the car can be perceived as a Trojan horse for the post-fossil recultivation of planet earth, the transition towards a solar culture and the decarbonization of the energy flows of the societal metabolism 2 (Rammler, 2011, p. 16). 2.4 E-mobility as a Sustainable System Innovation 3 Traps 2 Original quote in German: [Aus dieser Blickrichtung könnte man die Elektrifizierung des Bedürfnisfeldes Mobilität als eine Art] Trojanisches Pferd der grundsätzlichen postfossilen Rekultivierung des Planeten Erde, des Umbaus der fossilen zur solaren Kultur und der Dekarbonisierung der Energieflüsse des gesellschaftlichen Organismus betrachten. 35

49 In sum, when looking at these aspects related to concepts and visions of sustainable e- mobility, it becomes obvious that a focus limited to the diffusion of electric vehicles is too narrow in the context of sustainability transitions. A perspective on e-mobility as a sustainable system innovation would focus on the BEV as the central technology but pertain to a more systemic change in infrastructure, business models, behavioral patterns and, more generally, a different perception of the role of the car. Including a sustainability perspective, a transition towards sustainable e-mobility would be characterized by more integrated and intermodal forms of mobility, concepts of collective ownership of cars or carsharing and renewable energies as the basic electricity source. The question then is whether actual potential for this type of e-mobility going beyond the mere diffusion of BEVs can be observed and how to approach this analytically. The concept of system innovation captures this type of socio-technical change and is defined as a change from one socio-technical system to another. It comprises three aspects: One aspect of a system innovation is technological substitution, which comprises three sub-processes: (i) emergence of new technologies, (ii) diffusion of new technologies, (iii) replacement of old by new technology. The second aspect is coevolution. System innovations not only involve technological substitutions, but also changes in elements such as user practices, regulation, industrial networks, infrastructure, and cultural meaning. The third aspect is the emergence of new functionalities. When radical innovations have particular technical properties, this may enable the articulation of new functional characteristics. Radical innovations may then introduce new functionalities and change the way in which performance is measured (Geels, 2004a, p. 19 f.). This type of definition is especially useful when assessing transitions that have already been completed and can be studied ex-post, as historical cases (see for instance Geels, 2002, 2006a, 2006b, 2007). What about a case such as e-mobility, where various initiatives are currently on-going and at best indicate that we may be in an early stage of a wider transition? Hoogma et al. (2002) have argued that in these cases it is hardly possible to distinguish between innovations that will contribute to an optimization of established systems and others that will lead to substantial change but that nonetheless an effort should be made to determine the potential for radical change, depending on the specific technology and its context. This would require making an assessment, for each specific technology, of the likelihood that it could lead to co-evolutionary dynamics (Hoogma et al., 2002, p. 36). Looking at developments in the field of e-mobility, it is an open question whether there is a potential for the current momentum around BEVs to evolve as a sustainable system innovation. In order to focus on this question, three traps of e-mobility as a sustainable system innovation will be introduced as a way of guiding the analysis in the following. These traps are derived from the basic characteristics of a sustainable system innovation. According to the 36

50 definition by Geels (2004a, p. 19), a system innovation is characterized by (1) the diffusion of a new technology and (2) by the emergence of new functionalities as the result of sociotechnical co-evolution. In addition to these two aspects, a sustainable system innovation also needs to fulfill some sort of (3) sustainability criteria. When looking at the electrification of the car, these basic characteristics should serve as a heuristic for assessing the potential for a sustainable system innovation and can be reformulated in terms of traps that need to be avoided or surpassed. The first trap, the quantitative trap, would be that there is no sufficient diffusion of BEVs analogous to the first aspect of a system innovation: emergence and diffusion of a new technology. Obviously, diffusion is the basic process required for any technological innovation. Thus, it needs to be assessed whether there is sufficient diffusion and what might be particular barriers or driving forces. At the same time, it is clear that diffusion is more difficult for radical innovations that very clearly point towards a new system (Kern & Howlett, 2009, p. 394). The second trap, the qualitative trap, would be that there are no new functionalities emerging analogous to the second aspect of a system innovation: the emergence of new functionalities as a result of socio-technical co-evolution. Whether an innovation leads to a fundamental change depends on whether or not such new functionalities emerge. These result from the interplay of the new technology with other elements in their social environment, such as infrastructures, use patterns and business models, thus, co-evolutionary dynamics. It should be assessed whether patterns of co-evolution can be observed that involve new functionalities, i.e. re-defining needs and purposes, different ways of fulfilling needs and new criteria for measuring performance of a technology or a system as a whole. In the case of e- mobility, this would include new ways of fulfilling mobility needs, e.g. in an intermodal transport system where the role of the (electric) car has changed from the principal means of transport to an element in an integrated system guaranteeing mobility. Such a new functionality, i.e. a change in how mobility is provided, may also emerge at the interface of the transport and energy system and change the quality criteria of how well a system or a technological artifact performs: electric vehicles may not only be evaluated regarding their performance as a means of transport but also as a means of energy storage and how it integrates both functions. Both cases show that new functionalities emerging in the context of e-mobility will most likely imply changed perceptions of the role of the car in general. The third trap, the sustainability trap, would be that the new system configuration does not deliver sustainability benefits. This sustainability trap is included to deal with the 37

51 fact that sustainability is a central premise or background of research questions in this field of study, while it often is not further reflected in its concrete and case-specific implications, obviously due to the concept s inherent uncertainties and ambiguities. However, for the case of e-mobility, dealing with this sustainability trap would mean including concrete and casespecific sustainability criteria (e.g. renewable energies as electricity source, intermodality, carsharing). Fig. 4: Three traps of a sustainable system innovation. When e-mobility is analyzed from the point of view of a potential future system innovation that might eventually contribute to sustainable mobility, it is helpful to assess early on whether current developments point in this direction and according to what criteria this can be determined. It is also important to note that the three traps that might preclude the emergence of a system innovation are interlinked in various ways. Looking at the process of diffusion, it is obvious that electric vehicles are still developed and operated in small and protected niches. They have to be, because they cannot compete with conventional cars as regards range and price. Thus, in order to foster diffusion, electric vehicles are tested in niche segments, such as publicly funded pilot or demonstration projects, municipal fleets or carsharing systems. Already at this point even though critical levels of diffusion have not yet materialized signs of the emergence of new functionalities can be observed. In demonstration projects, public funding facilitates the building up of charging infrastructures creating the link between e-mobility and the energy system, which may induce 38

52 a new perception of cars not just as means of transportation, but also electricity storage capacity. Corporate R&D projects create niches where innovative business models are tested, such as carsharing, battery leasing or battery-swap services, which may eventually pave the way towards new forms of mobility that do not solely rely on private car ownership (Tran et al., 2012, p. 332). Test users in demonstration projects experience electric vehicles in everyday social practices and it can be shown empirically that use patterns are adapted to the new technology while it is actually used and often in ways that are different than had been assumed beforehand (Canzler & Knie, 2011, p. 42). Simultaneously, these processes may have an impact on diffusion rates. In science and technology studies as well as in technological sociology, it has been shown that technological change is always an inherently social and cultural process (cf. Bijker et al., 1987; Dijk, 2011). Whether a new technology becomes established depends on the cultural significance that is assigned to it. It can be successful if meaning is connected to it and if it is embedded in its social context. This can be achieved by proactive advocacy of influential societal groups or actors, or it may emerge through co-evolutionary processes linking a new technology with wider societal developments (Weider & Faul, 2011, p. 100 f.). Thus, developing electric vehicles in niches that give them a specific meaning (e.g. as an element in the energy transition) and link it with societal trends (e.g. young people living in cities that use carsharing services rather than buy a car) increases the chances for diffusion. These examples show that socio-technical co-evolution plays a central role, because it may lead to the emergence of new functionalities (i.e. in fact a redefinition of the car). New forms of sustainable mobility are thus much rather the result of socio-cultural changes as a response to technological innovations, than of technological innovation alone (WBGU, 2011, p. 264 f.). Especially the electric vehicle as a radical innovation has the potential to initiate a process of redefining current conceptualizations of mobility (Vergragt & Brown, 2007, p. 1108; Vallée & Schnettler, 2013, p. 34 f.; Weider & Rammler, 2011, p. 5). From a historical point of view, Weider and Rammler (2011) argue that the pioneering element in Carl Benz s work was not per se the invention of the internal combustion engine, but rather the creation of the automobile principle, the ability to move autonomously as an individual (p. 3). However, as with many innovations, it is not clear whether the electric vehicle will evolve in the context of a wider system innovation, persist forever as a niche innovation, or vanish again completely. It is relatively certain that it will not replace the conventional car without any other changes, but there is still a variety of options ranging from e-mobility as a system innovation to electric vehicles being used primarily as a household s second private car or in 39

53 small market segments, such as urban fleets or carsharing services (Kemp & Rotmans, 2005, p. 35). These examples highlight the importance of considering the third trap, referring to the potentially unsustainable outcome of a system innovation, already at an early stage. The case of e-mobility clearly shows that a focus on the development and diffusion of electric vehicles does not necessarily produce sustainable results. It is important to adopt a systemic perspective and assess whether the emergence of specific types of new functionalities point to overall more sustainable results. The principles of sustainable mobility discussed in section 2.2 provide an orientation, emphasizing that an overall reduction in the volume of transport or shifts to other modes of transport are important elements. This means that it is, for instance, important to consider whether developing electric vehicles in niches eventually aims at creating a market for selling and using them in the same way as conventional cars, or whether carsharing and intermodal transport schemes lead to an ecologically beneficial modal shift instead of additional demand for mobility. Analytically, an attempt should be made to assess whether activities in fostering e-mobility provide an alternative to current unsustainable patterns in fundamental ways, in the sense that whatever future technological paradigms emerge over time in the future should be set in a post-hydrocarbon landscape (Cooke, 2011, p. 107). In sum, it is suggested that including the three traps can serve as a helpful analytical lense for the study of cases such as the electrification of the car, which are at best at an early stage of a wider transition. The three traps should in this way serve as a guiding concept for assessing system-innovative potential. The obvious difficulty of such an endeavor should not lead to neglecting these issues. 2.5 Current State of Research: Literature Review A literature review has been carried out guided by the question of whether there are fields of research or theoretical approaches that focus on e-mobility and relate it to broader definitions of sustainability. The three traps are used to guide and systematize the results of the literature review, which aims to show whether approaches exist that can really capture the complex phenomenon of e-mobility as a (potential) sustainable system innovation. Therefore, literatures are searched that focus on e-mobility as defined in this thesis (ch. 2.1), i.e. centered on electric vehicles as a truly radical innovation as compared to hybrids or fuel cell vehicles, but not limited to electric cars. Various types of electric vehicles are considered relevant when e-mobility is understood as a systemic type of change in mobility in general. This also means 40

54 that literatures are included that are dealing with issues of sustainability mobility primarily, but discuss e-mobility as well to at least some extent. The method applied is a database search (conducted in July 2013) by keywords and is limited to peer-reviewed journal articles. It is further limited by excluding the purely technological disciplines (mechanical engineering, electrical engineering, chemistry) from the substantive analysis, because they exclusively focus on technological solutions. These literatures are only assessed, in order to identify the quantitative share of the exclusively technological aspects in the field as a whole. Other relevant literatures are transport and innovation studies as well as economics and business approaches focusing on the marketability of electric cars, new business models, the willingness of consumers to buy electric cars and the impact of this new technology on industry structures and national economies. Finally, the field of sustainability transitions research will be included as well. These literatures will be searched by a combination of keywords on e-mobility (electric cars, battery-electric vehicles etc.), sustainability as well as keywords from the field of sustainable mobility in general, as identified in the respective literature discussed in ch. 2.2 (e.g. Banister, 2008; Cohen, 2006). The results of the database search will be analyzed and publications will be systematized according to the extent that they address the three traps of a sustainable system innovation. A search for articles in the Web of Science using the keywords e-mobility*electric mobility produced roughly 300 results. In order to exclude articles that are not related to e- mobility in transport (such as biochemistry, physics etc.) the term electric vehicle was included, which decreased the number of results to 148 articles. Screening the titles and abstracts of these 148 articles led to the manual exclusion of another 74 articles not related to the topic of electric automobility (but e.g. articles from the fields of electronics and communication technologies, medical science, chemistry, physics, biochemistry) as well as articles from the field of transport with a focus on biofuels or electronic traffic management systems, articles that focus exclusively on battery technology and only mention electric vehicles as one field of application, and articles that deal with specific military applications of e-mobility. Thus, a sample of 74 papers has been identified in this first step. As a second step, the search was refined by including the terms sustainability*transition*system innovation, in order to identify publications that focus on issues of sustainability and systemic types of change around the electric vehicle. This step reduced the original sample of 148 papers to 20 articles. From these 20, 8 articles from the fields of biochemistry and physics were manually excluded (as in step 1). Thus, 12 papers 41

55 from the sample of overall 74 papers have been identified to be particularly relevant with regard to the focus and research question of this thesis. As a third step, the Web of Science was searched by the terms mobility*transport/strategic niche management*technological innovation systems*multilevel perspective*transition management, thus the different theoretical approaches in the field of transition studies (see ch. 3.1), in order to provide an overview of transition papers in the field of mobility in general. This step produced a result of only five papers, all of which had already appeared in the original search. One of these five papers (Berkhout et al., 2010) had already been excluded from the earlier sample, because it has no specific focus on (e- )mobility. Another paper by Hellsmark & Jacobsson (2009) is excluded because it deals exclusively with biogas as an alternative fuel. Thus, this step produced another three papers from the sample of 74 papers, which can be considered among the relevant papers, now counting 15 articles. As a fourth step, the search terms used in step three were also used to perform a search in Google scholar, since many contributions from the still young field of transition studies are not (yet) listed in the Web of Science. This can be explained by the fact that transition studies is still a newly emerging field of research with a weak citation basis which may be at least part of the explanation. The search in Google scholar (same search terms as in step 3) delivered more than 1,000 results. Scanning the first ten pages of search results, 13 papers from the field of transition studies were identified. This was done by identifying topical keywords, journals and authors that can be grouped among the field of transition studies according to the survey of the field of transition studies by Markard et al. (2012). Of these 13 papers, two had already turned up in the Web of Science search (Kemp et al. (1998), Köhler et al. (2009)) and have already been included among the more relevant articles. After this step, the total sample has increased to 85 articles and the more relevant among these are now 26 articles. Finally, the journal Environmental Innovation and Societal Transitions as a new journal, established from within the field of transition studies has been scanned for relevant publications. This adds to the preceding Google scholar search and was done manually, because the journal is not appearing in the Web of Science yet. Two relevant contributions were identified, one of which had already turned up in the Google scholar search. Thus in sum, the literature review produced a sample of 86 articles, of which 27 have been identified as particularly relevant with regard to a focus on e-mobility as a sustainable system innovation. With regard to the distribution across disciplinary fields, it can be shown 42

56 that a large share of the overall sample, i.e. 34 articles, can be assigned to the field of energyrelated research. The majority of papers deal with e-mobility from an energy systems perspective, focusing on issues of integrating electric vehicles in energy infrastructures. The second-largest group of 20 articles can be associated with engineering disciplines, and the remaining papers can be grouped across transport studies (14 articles), environmental impact or life-cycle analysis (7 articles), and the field of transition studies (11 articles). This distribution pattern is depicted in Fig. 5. Fig. 5: Distribution of papers across fields of research. Fig. 6: Topical focus of papers. By scanning the abstracts of the 86 articles for basic research questions, research design and results, an overview has been gained with regard to the basic topical focus of the different articles. Fig. 6 shows that a majority of the articles depart from either an economic or market perspective focusing on issues of diffusion of electric vehicles, user acceptance and industry dynamics, or from an energy system perspective. Issues of vehicle technology and environmental impact assessment of electric vehicles form another relatively large group. Policy-related aspects make up a small amount of the sampled articles and these focus predominantly on policies fostering innovation and diffusion processes of electric vehicles. A more integrated perspective on e-mobility can be found in only a few articles, mostly 43

57 published in journals that can be associated with the field of sustainability transitions research. A more detailed look has been taken at the 27 articles identified as particularly relevant in the context of sustainable e-mobility and system innovations. An overview of these articles is presented in Tbl

58 Tbl. 1: Overview of papers with a focus on e-mobility as a sustainable system innovation. 45

59 These 27 articles have been analyzed with regard to their basic problem perceptions and the scope of their research approach in terms of the three traps of a sustainable system innovation developed in this thesis. It has been assessed whether these articles predominantly focus on issues of diffusion of electric vehicles ( the quantitative trap ), socio-technical coevolution ( the qualitative trap ), and whether they include an explicit perspective on sustainability ( the sustainability trap ). It has also been assessed how broad or narrow their view on sustainability is and if some of them possibly also provide an integrated perspective on sustainable system innovations, thus addressing all three traps. An overview of the 27 articles and the distribution of articles regarding their scope in terms of the three traps is presented in Fig. 7. The more in-depth screening of the articles led to the exclusion of 5 articles, which provide more general conceptual perspectives on sustainable mobility with e- mobility being not more than a side note or just a very brief example (Geels, 2012; Geerlings et al., 2009; Kemp et al., 1998; Köhler et al., 2009; Schot & Geels, 2008). Even though most of the remaining 22 articles adopt an integrated perspective of some sort, for instance by clearly positioning e-mobility as an environmental innovation (Oltra & Maïder, 2009), a majority, i.e. 12 articles, in essence deal with the problem of diffusion of electric vehicles and focus on industry and market dynamics, business strategies, or suitable policy instruments. For instance, Weber and Hoogma (1998) study different styles of innovation diffusion and emphasize that this approach takes into account cultural aspects of management traditions, social interpretations of technologies and the influence of local circumstances (p. 545). However, in their analysis the social elements in socio-technical coevolution are not dealt with explicitly and the aim of the paper is to examine variables influencing the diffusion of innovations, thus remains limited to the first trap. Electric vehicles are predominantly discussed in terms of a technological substitute: A high proportion of trips made for commuting, shopping etc. are short and could, in principle, be substituted by EVs (p. 558). The question of whether and when diffusion processes may result in radical innovations is pondered upon only briefly. Similarly, in a more recent article by Sierzchula et al. (2012), the basic research question also deals with the likelihood of EVs eventually becoming a legitimate competitor to ICE vehicles (p. 51). The aim is to understand the scope of the current transition (ibid.), while this scope is envisioned to range from electric vehicles as a substitute for conventional cars at one end of the scale, to electric vehicles remaining a niche technology at the other. Thus, a transition is conceptualized as a diffusion challenge rather than a more fundamental system innovation resulting in new functionalities. There is no rethinking of the basic scale for assessing the possibilities of e- 46

60 mobility and it remains limited to a market penetration framework: what is measured is a difference in degree (of diffusion), while differences in kind (system innovation) are outside the framework. A similar focus on innovation dynamics and developments in e-mobility as a technological niche can be found in the articles by Bakker et al. (2012), Hasegawa (2010), Magnusson and Berggren (2011), and Kühne (2010). Apart from a focus on industry dynamics, technological development and related business strategies, the remaining papers that can be characterized as addressing predominantly the first trap study issues of user acceptance in the context of sustainable lifestyle practices (Axsen et al., 2012) or deal with the question of whether electric vehicles are an attractive option for customers (Garcia & Miguel, 2012). Another three papers in this group add to industry and market perspectives by introducing a focus on governance and, in particular, policies for supporting innovation and diffusion of electric vehicles. For instance, the paper by Åhman (2006) discusses the role of political support for technological development and early market introduction of alternative drive technologies in Japan. It adopts a systemic perspective insofar as it considers the whole policy context identified to be relevant, instead of an individual policy instrument, as well as other framework conditions, e.g. business attitudes and broader institutional settings (p. 434). However, the approach is not systemic in terms of the three traps, since it focuses on technological substitution and diffusion processes, assessing the chances of a technology surviving the long journey from idea to competitive technology (p. 442). The papers by Browne et al. (2012) and Nilsson et al. (2012) depart from a concern for environmental innovation in the transport sector and also employ a systemic perspective on governance approaches and patterns. However, policies are analyzed with regard to their ability to foster the diffusion of alternative drive technologies and, for instance, major barriers are identified to be technical limitations, commercial feasibility and market availability (Browne et al., 2012, p. 150), thus obstacles for technological substitution, rather than system innovation. A broader scope and focus on socio-technical co-evolution can be identified for four articles. They look beyond market-based diffusion and technological development and may thus contribute to understanding the emergence of new functionalities, i.e. addressing the second qualitative trap. For instance, Dijk et al. (2013) study the emergence of an electric mobility trajectory, which is the result of interrelated technological and social dynamics, or, van Bree et al. (2010) analyze how the relationship between car manufacturers and consumers is influenced by socio-technical co-evolution in the field of hydrogen vehicles and e-mobility. The fundamental difference in these approaches, as for instance compared to Browne et al. 47

61 (2012) focusing on technological and market immaturities, can be shown by what are identified as crucial factors in the development of e-mobility: developments in (1) infrastructure, (2) developments in mobility, (3) developments in the global car manufacturing regime, (4) developments in energy prices, and (5) developments in the electricity sector (Dijk et al., 2013, p. 140). Similarly, Dijk and Yarime (2010) explicitly focus on the role of techno-economic mechanisms alongside social and regulatory mechanisms (including the social meaning of an engine) (p. 1371) in their study of the emergence of hybrid-electric cars. Cultural connotations and meaning of technological artifacts such as cars or car engines are an important element for understanding the concept of new functionality resulting from processes of co-evolution and they are, in fact, the core of a system innovation. A more concrete case of new functionalities emerging in the field of e- mobility is presented by Falvo et al. (2011): an integrated urban mobility system, linking metro transit systems and electric vehicles via smart grids. Even though this article is more technology-oriented in nature, it includes a broad perspective on urban infrastructures and it implicitly points towards new functionalities emerging at the interface of public and private transport and the energy system. Even though sustainability is mentioned as one basic rationale for developing e- mobility and a motivation for the research of most papers, it is more explicitly addressed and a central part of the analysis in only four articles. Life-cycle analyses are conducted to assess the environmental impact of electric vehicles depending on different energy sources, i.e. electricity mixes, for powering vehicles and comparing them to other types of alternative drive technologies (Bartolozzi et al., 2012; Faria et al., 2012). One article focuses on environmental management systems and ISO certification processes for hybrid electric vehicles (Orecchini & Sabatini, 2003) and the paper by Stamp et al. (2012) assesses the impact of different supply options for lithium carbonate and how this affects the overall ecological balance of electric vehicles. These articles explicitly focus on environmental sustainability concerns related to electric vehicles and therefore they do address the third trap, which is crucial when considering the heated debates about whether e-mobility can actually provide solutions to environmental problems in the transport sector. However, they do not integrate more systemic perspectives on sustainability, which are relevant in the context of sustainable system innovations. The remaining two papers have been found to address all three traps, albeit in slightly different ways. In their case study of the company Better Place, Christensen et al. (2012) focus on socio-technical co-evolution in terms of a new business model developing around 48

62 the electric vehicle as a new technology. The potential for widespread diffusion as well as related environmental impacts are reflected upon. The basic question of developing innovative business models for e-mobility is placed in a broad context, addressing the problem of diffusion, arguing that innovative business models may be a prerequisite to the success of new technologies in the market (p. 499), of socio-technical co-evolution and the emergence of new functionalities by focusing on the need for fundamentally different business models and new forms of convergence between the automotive and the energy sector, and finally, of (environmental) sustainability as the second condition apart from market breakthrough determining the success of developing e-mobility. It is further shown that these aspects are interrelated: If the model remains small, with just a few hundred BEVs in circulation, then it is likely to fail as a business but also will be insignificant as a contribution to more sustainable mobility (p. 504). The second paper by Nykvist and Whitmarsh (2008) studies sustainable mobility transitions where technological innovations, e.g. electric vehicles, are one element of such a transition among others, such as modal shift and reduced travel demand. The argument is based on a broad definition of sustainability including not only environmental protection but also social and economic welfare (p. 1373). Diffusion of technology is addressed in terms of niche development (p f.), which is however not limited to technological niche innovations, socio-technical co-evolution is addressed by exploring institutional, behavioral and technological components of niche development (p. 1375), and these issues are analyzed against the background of a broader sustainability transition in the field of transport (p. 1348). 49

63 Fig. 7: Distribution of selected papers across three traps. In sum, the literature review has shown that research on e-mobility is dominated by disciplinary perspectives on technological and economic issues relevant for the diffusion of electric vehicles. Even in transition studies the focus often is on questions of greening technologies and diffusion, or on questions of adapting the socio-economic systems and political institutions around presumably green technologies (such as the electric vehicle). What is rarely found is a clear perspective on socio-technical co-evolution (How do electric vehicles trigger change in use patterns or business models? Can patterns be observed of how new functionalities do actually emerge?). Furthermore, a basic concern for sustainability appears throughout all of the papers and it is mentioned as the context in which electric vehicles are developed and studied. A more pronounced perspective on sustainability can be found in life-cycle assessments, but these usually remain limited to a narrow ecological dimension of sustainability. A broader perspective on sustainable mobility, assessing the degree to which electrified forms of mobility are potentially more sustainable, not only with regard to emissions and resource depletion but also including social aspects, is lacking. Two 50

64 articles have been identified that are addressing all aspects of a sustainable system innovation. However, the paper by Christensen et al. (2012) employs a narrow environmental definition of sustainability, while the paper by Nykvist and Whitmarsh (2008) departs from a broad conceptualization of sustainability, but focuses on sustainable mobility in general where e- mobility is one aspect among others. This implies that e-mobility is a field of research that is dominated by economic and technological approaches, possibly due to the fact that the electric vehicle as a technological innovation is central and currently high on public agendas and discourses in the context of economic potential for national industries. A parallel and largely unconnected research strand focused on sustainable mobility in general seems to exist that is reluctant to engage in a field dominated by market and technological perspectives, or does not realize that there are some crucial interlinkages. Seeing that e-mobility currently finds broad political support, at least partly out of a concern for more sustainable transport systems, it seems promising to focus on and re-evaluate e-mobility from the point of view of broader sustainable mobility and system innovation perspectives. A similar assessment with regard to the state of research has been made by Lyons (2012), showing that in transport studies two sharply separated strands can be identified, either focusing on behavior change, or on technological fixes in the context of sustainable transport. In order to deal with current challenges in the transport sector, Lyons (2012) suggests that we have yet to embrace a socio-technical transition perspective that seeks to understand how technology and the behavior change of all actors (individuals, firms, policymakers etc.) dynamically interact and co-evolve (p. 30). Another review of the literatures on climate change mitigation in transport has been presented by Schwanen et al. (2011). Here, it is also shown that economic and technological approaches are dominating this field of study and it is suggested that contributions from the social sciences, amongst others from the field of transition studies, could be useful for a more profound understanding of the sources for unsustainability in the transport sector and for providing more comprehensive approaches to solving current challenges (p. 933). This implies that in this particular field any future research agenda needs therefore to understand more fully how embedded practices and technological change inter-relate in specific concrete conditions (Wells & Nieuwenhuis, 2012, p. 1681). One aim of this thesis is to close this research gap and as a next step the literature on system innovations and sustainability transitions will be discussed as a possible framework for better understanding the potential e-mobility could develop in the context of more sustainable forms of mobility. 51

65 3 A THEORETICAL FRAMEWORK FOR ANALYZING SYSTEM INNOVATION Transition studies, or the field of sustainability transitions research, explicitly deals with sustainability-oriented system innovations, thus the paradigmatic change hoped for when envisioning future e-mobility. The aim of this chapter is to develop a theoretical framework for analyzing the potential of e-mobility to emerge as a sustainable system innovation drawing on concepts from the field of transition studies. A further aim is to position the thesis in the context of this particular field of research. The chapter is structured as follows. An introduction and overview of Sustainability Transitions Research as a distinct field of research is presented in chapter 3.1, including an overview of basic definitions, their origins and the different approaches for analyzing transitions (ch ), and the specific role of technology and innovation in this context (ch ). The multi-level perspective on transitions as the central theoretical approach in this field will be discussed in detail (ch ), including a critical reflection on its scientific value, its shortcomings and applicability to the research question posed in this thesis (ch ). A major challenge is the question of how to deal with cases of transitions in the making, such as e-mobility where it is currently not clear whether developments in this field will eventually amount to a transition or system innovation contributing to future sustainable development. The aim of this thesis is to find a way of analyzing the potential inherent in currently observable dynamics in the light of an envisaged sustainable system innovation. Therefore, ch. 3.2 is dedicated to the issue of analyzing such transitions in the making. Subsequently, a theoretical framework building on the multi-level perspective and addressing some of its weaknesses is developed in chapter 3.3. Building on a discussion of the relevant unit of analysis in studying system innovations (ch ) and a specific conceptualization of system innovation (ch ), Dolata s (2009, 2011, 2013) concepts of transformative capacity of a new technology (in this case: do electric vehicles trigger social innovations, e.g. new business models, political regulations or use patterns?) and adaptability (here: how inert or open to change is the mobility system?) are introduced in chapters and It will be shown that including these concepts into an overarching multi-level perspective on transitions provides a way of analyzing the basic dynamics of a system innovation on the micro-level and give an indication of its actual potential in an ex ante situation (ch ). In short, the aim of the following chapters is to arrive at a theoretical approach that transcends the logic of incremental innovation and clearly conceptualizes processes of systemic socio-technical change. 52

66 3.1 Sustainability Transitions Research The field of sustainability transitions research builds on innovation studies, science and technology studies, complex systems and innovation systems theory as well as concepts from sociology and evolutionary economics to study fundamental and long-term change in sociotechnical systems, such as transport, energy, water and agri-food systems (van den Bergh et al., 2011). The basic motivation tying together scientists from these different disciplinary backgrounds focusing on different empirical domains is a concern for sustainability. The shared observation is that current environmental problems related to climate change and resource depletion are hardly being solved and that a better understanding of the reasons for their persistence and more comprehensive approaches for addressing them are needed. The basic problem, which is explicitly addressed in transition studies, is that sustainability problems or causes for unsustainability are deeply rooted in modern societal structures and thus represent the dark side of dominant patterns of socio-economictechnological development (Grin et al., 2010, p. 2). The sectors or socio-technical fields that are most relevant for the functioning of modern societies, e.g. energy, transport and food production, are the locus where change is needed most desperately and which at the same time are characterized by technological lock-ins and path dependency that explain their historical success and now preclude any substantial change. In any of these fields, dynamic stability arises from the many interlinkages and interdependencies between dominant technologies, existing infrastructures, political and societal institutions, markets and business models as well as lifestyles, cultural meaning and individual behavior. In these interlinked and complex configurations that have historically grown into relatively stable systems, change is usually limited to incremental innovation aimed at optimizing and preserving (and thus not threatening power positions and vested interests of central actor groups). This makes sense in so far as a specific system has proven to be successful over time and thus is willfully being sustained. However, since current environmental problems are severe and directly caused by basic structural conditions and principles of today s system configurations, they cannot be solved by traditional means and strategies aimed at, basically, preserving current systems. Thus, a more radical type of change is needed that will possibly shake the basic foundations of modern energy or transport systems. This type of change is captured by the term transition, highlighting that incremental changes will not suffice to cope with the prevailing sustainability challenges (Markard et al., 2012, p. 955). This requires new approaches for integrated socio-technical change, that go beyond the development and diffusion of 53

67 environmental innovations such as green technologies, and which also include more comprehensive policy approaches (Grin et al., 2010, p. 2 f.; Markard et al., 2012, p. 955; van den Bergh et al., 2011, p. 8). The whole transition endeavor is guided (often even only implicitly) by a very broad notion of sustainability. Sustainability is the basic motivation for engaging in transition research and it serves as an orientation for concrete cases. In general, sustainable development is understood to be an open-ended search process, rather than a clear future vision. Assessments of specific policies, technologies or broader patterns of change with regard to their contribution to overall sustainable development are to be carried out on a case-by-case basis. This approach is presented as a specific strength of transition studies, because the basic tenets of the sustainability debate (based on the Brundtland definition and ensuing discourses) provide orientation, while concrete goals and strategies remain open to political negotiation, as is appropriate with regard to the many ambiguities inherent in the idea of sustainability (Grin et al., 2010, p. 2). However, it is important to be aware of the fact that implicit understandings of sustainability can be very different with regard to images of a desirable future. They may range from visions of a green economy to concepts that question whether it is at all feasible to reconcile traditional market economies with principles of sustainability (cf. Kallis, 2011; van den Bergh, 2011; Victor, 2010). Such differing conceptualizations of sustainability have consequences for the notion of transitions towards sustainability that is used in different contexts. The term transition can mean fundamentally different things depending on whether it is applied in varying science or policy contexts, or, for instance, in the transition towns movement (Hopkins, 2011). Thus, careful reading of what the underlying assumptions are of different usages of the term transition is imperative and it is important to note the varying implications they have with regard to power structures in specific socio-technical systems or modern societies in general. It can be argued that most transition studies in the specific field of research in focus here, do not fundamentally question basic premises of market economies and the need for economic growth. Based on the insight that many sustainability problems are directly linked to basic societal structures, it is advisable that the promises of economic growth and technological innovation as a means to foster growth and reconcile it with environmental protection should at least be scrutinized critically (Stirling, 2011; van den Bergh et al., 2011, p. 14). Having pointed to these different contexts where the notion of transition plays a role, the specific field of sustainability transitions research, which is in focus here, has emerged 54

68 from various disciplinary backgrounds and research traditions and a distinct scientific community has begun to emerge during the late 1990s. A bibliographical analysis carried out by Markard et al. (2012) to capture Sustainability Transitions as an emerging field of research shows the dynamic development of this still young scientific community. Up until today, more than 500 publications can be assigned to the field and more than 6,000 citations until the beginning of Important thresholds in the development have been identified around the year 2005 when the number of yearly publications began to increase significantly and subsequent increases showed in the number of citations per year, from 235 in 2006 to 1815 in 2011 (Markard et al., 2012, p. 960 f.). This increased interest in the field during the early 2000s can partly be explained by developments in the Netherlands at that time. Political efforts by the Dutch Ministry of Environmental Affairs in the field of sustainable energy were developed under the heading of transition management, developed together with scientists. Innovative approaches were introduced in the policy process, including politicians, researchers and other stakeholders, especially from the energy industry. The notion of transition thus spilled over to the political realm, which in turn had the effect of motivating further efforts by researchers to engage in this field (Chappin & Ligtvoet, 2014, p. 716). This particular role played by the Dutch political context may also explain the empirical focus emerging within transition studies. Looking at scientific publications in this field again, most of them focus on energy-related topics, followed by transport, water, and food production. Furthermore, most of the publications focus on empirical cases in the Netherlands and the UK, followed by the US and Germany (Markard et al., 2012, p. 961). Even though the field of transition studies has grown significantly, it is still represented by a relatively small scientific community that can be characterized as more or less homogeneous as regards the focus of research strongly linked to the Dutch energy policy context and building on a relatively limited number of key references as well as the tightly interlinked networks of co-authorship. The field has also been formally institutionalized by holding yearly conferences, establishing an official network (Sustainability Transitions Research Network, STRN, circulating a regular newsletter, and issuing a scientific journal specifically dedicated to transition studies ( Environmental Innovation and Societal Transitions ) (Farla et al., 2012, p. 991; Markard et al., 2012, p. 956). While all this may be considered a normal and necessary consolidation process within a young field of research, some critical thoughts have been voiced, especially with regard to homogeneous patterns of co-authorship and cross-referencing each other s work. Openness 55

69 with regard to approaches in other scientific communities and an awareness of existing theories and bodies of literature can help avoid a tendency towards unduly limiting the perspectives within transition studies (Chappin & Ligtvoet, 2014, p. 720 f.). Similarly, critical voices have pointed out that transition may be just a new buzzword for other, much older concepts. In fact, transition studies have emerged from various disciplinary backgrounds where similar notions and concepts pre-exist. For instance, in economics and in innovation studies, concepts of radical innovation and industrial transformation can be found. Structural change and transformation of economies or societies have also been dealt with in political science and sociology. Also, within interdisciplinary fields of environmental or sustainability research, concepts of ecological modernization have been developed. However, it is pointed out by proponents of sustainability transitions research as a distinct field of research, that none of these approaches can alone grasp all the dimensions that are relevant when looking at current environmental problems in broad sociotechnical fields, such as energy or transport, while such a comprehensive perspective is needed, in order to address these complex real-world problems. Obviously, the problems in the transport sector cannot be meaningfully addressed by only focusing on technological innovation, or on market dynamics, or on political regulation, or on changes in consumer behavior. The basic rationale for a new approach of sustainability transitions research is to provide a perspective where these different elements and their interlinkages can be grasped in a comprehensive way (Geels, 2012, p. 471 f.; Geels, 2011, p. 38; van den Bergh et al., 2011, p. 8). Based on these considerations, it can be argued that the field of sustainability transitions research has its place, especially against the background of current and future sustainability challenges. However, the question remains, how exactly can transition approaches integrate all those different aspects in meaningful ways? This points to the dangers inherent in combining concepts and approaches from different research traditions and based on different ontological positions (Garud & Gehman, 2012, p. 980 f.). While some may relatively easily be integrated in a comprehensive framework, others may be more difficult to combine, because they may be based on opposing theoretical assumptions and rivaling explanations that cannot be reconciled. Thus, some caution is warranted to avoid matching together approaches that depart from opposing premises and basic assumptions. Nonetheless, comprehensive heuristic frameworks within transition studies can be helpful in fostering an integrated perspective across different academic disciplines, which is important for addressing complex sustainability problems. Thus, instead of avoiding integration altogether, the 56

70 diversity of ontological backgrounds and theoretical ambiguities should be made explicit and welcomed by reflecting on inherent pluralities and broadening perspectives (Stirling, 2011, p. 84 f.). Therefore, the basic concepts in the field of transition studies, their definitions and their ontological backgrounds will be spelled out in the following Central Concepts and their Origins The notion of transition is the central concept around which a distinct field of research has emerged. The term is defined broadly as a process of fundamental change in a socio-technical system, which provides a specific function or service for society as a whole, e.g. energy, mobility, food or water and sanitation. Substantial change within these systems is necessarily radical or of a fundamental nature, because these systems are complex as regards their actor constellations, including governmental actors, industry and firms, civil society organizations and individuals, they are built around large infrastructures and specific technologies and their system boundaries do not coincide with typical disciplinary, sectoral or institutional boundaries. Rather, a socio-technical system is delineated based on the specific network of actors, technological infrastructure and institutions (formal, such as regulations, and informal ones, such as guiding principles and norms) that together fulfill a specific societal function in a characteristic way. In this way, the concept of socio-technical system in transition studies can be distinguished from the way it is used in organization studies, where it originated (cf. Trist & Bamforth, 1951; Trist, 1981), not with regard to its basic principles though, but with regard to scope and field of application. A full-fledged transition in such a socio-technical system thus means structural change in various societal domains and it can be expected to span at least one generation, thus a period of 25 to 30 years. The long-term and complex nature of transitions implies a process of multiple decisions in different arenas across an extended period of time, thus including feedback loops and constantly changing circumstances. In the end, new actors and networks, new business models and consumption patterns will have emerged. The role of specific technologies changes as well as basic institutional structures and societal perspectives on the way that, for instance, energy or mobility are provided in general (Hoogma et al., 2002, p. 198; Markard et al., 2012, p. 956; Nilsson et al., 2012, p. 52 f.). The notion of transition has prominently emerged in current discourses on future sustainable development however, any meaningful empirical observation of a transition as an actual phenomenon can be made only with a view to past transitions, e.g. fundamental changes in transport systems at the turn of the 20 th century, the industrial revolution or other 57

71 major historical transformation processes. These transitions were emergent, as opposed to current efforts of transforming modern societies, with concrete initiatives and strategies being politically negotiated and decided, planned and mapped out where possible, in order to facilitate future sustainable development. Furthermore, the term transition is used as a guiding vision, while, strictly speaking, it is only possible to talk about a transition, once it has been completed and an ex-post perspective allows the observer to assess whether fundamental change has actually taken place. Thus, shifting the focus from transitions in general to sustainability transitions, means that the direction of an open-ended change process becomes a central element. In contrast to emergent transformation processes, sustainability transitions have an inherently normative component, insofar as they are focused on change towards more sustainable production and consumption patterns in modern societies. What sustainability means in concrete cases is certainly ambiguous and contested, thus generally subject to debate and conflict, and, what is more, concrete case-specific sustainability criteria may also change over time. Consequently, an important element in sustainability transitions is the way that these are guided and directed proactively (which is different from emergent transitions, not following an overarching agenda). Often, political actors and governmental regulation play an important role because sustainable development has become an important part of political strategies and agendas at different levels. However, since transitions of sociotechnical systems extend over different societal domains, other actors need to be involved, too, such as industry representatives, firms, consumers, researchers and civil society organizations, and they should cooperate in coordinated ways. This is particularly important in the case of sustainability transitions, because in most cases more sustainable solutions cannot compete on established markets and, as in any case related to public goods, offer benefits to society as a whole but not directly to individual firms or consumers. Thus, sustainability transitions rarely emerge smoothly from current system structures and usually depend on facilitating changes in political and economic framework conditions. This leads to another central element in sustainability transitions: the role of vested interests and power struggles. Actors with a vested interest in the status quo will utilize the ambiguous nature of sustainability and try to influence the debate in such a way that fundamental change is hampered or precluded altogether. A particularly important role is played by large incumbent firms that dominate socio-technical fields, such as transport or energy systems. Due to their influence and resources they could be important for fostering the breakthrough and acceleration of change and alternative solutions. However, since fundamental change to 58

72 current systems threatens their vested interests, they usually use their power to prevent change and secure their current positions (Geels, 2011, p. 25; Markard et al., 2012, p. 956 f.). Both transitions and sustainability transitions focus on change in socio-technical systems, defined by their provision of a basic service for society such as mobility or energy. Especially when interested in transitions to sustainability, defining the object of study in this way can help to analyze and address the fundamental causes of unsustainability. As already mentioned, sustainability problems can be characterized as persistent, because they are so tightly interlinked with evolved societal structures and thus the way societies fulfill their various specific functions and purposes. Transitions in socio-technical systems directly address these basic societal functions and purposes, thus the root causes of sustainability problems, rather than more isolated aspects such as, for instance, green innovation, environmental policy or sustainable consumption patterns. As a consequence, in transition studies, the unit of analysis is not clear-cut in empirical terms: important delineation criteria for a socio-technical system as object of study are motivated based on sustainability concerns and revolve around the concept of a generic societal function. Relevant systems can then range from national energy systems to healthcare systems, to regional entities such as cities or ecological systems, or more abstract units of analysis, such as political systems or financial systems (de Haan & Rotmans, 2011, p. 92). This means that typical units of analysis from a wide range of disciplines can be studied from a transitions perspective while the common denominator is the orientation along functions and real-world problems, which is decisive for dealing with sustainability problems adequately and require defining system boundaries that often deviate from disciplinary-based definitions in academia. It should be noted that such an approach requires the researcher to be explicit about the object of study and to clearly delineate the respective unit of analysis. A similar emphasis on generic functions and underlying structures is captured in the concept of system innovation, which is a dominant concept especially in earlier transition studies. The basic definition has already been introduced in chapter 3, laying out the three aspects of a system innovation: (1) technological substitution, (2) socio-technical co-evolution, and (3) the emergence of new functionalities (Geels, 2004a, p. 19 f.). 59

73 Much like the term transition, a system innovation is used to describe fundamental changes in a socio-technical system. The emergence of new functionalities that results from coevolutionary developments is the end-state of a transition where dynamics have interlinked across different levels and between societal domains, resulting in a new and different system configuration. The term new functionality is used to describe how a system innovation (or transition) leads to a qualitative change in the way that a specific socio-technical system fulfills its particular function. While providing mobility or energy are basic functions that any modern society will have to fulfill, a sustainability transition in these fields might lead to fundamental changes in how this is done (e.g. energy provision in a central or decentralized energy system), what performance or quality criteria are applied (e.g. access to mobility vs. ownership of a car), and regarding the actor structure and dominant technologies. In contrast to the term transition, the notion of system innovation considers more explicitly the role of technologies (which are central elements in many sustainability-related problems and their potential solutions), while it also emphasizes that substantial change is not purely technological, or triggered uni-directionally by technological innovation, but rather the result of socio-technical co-evolution (Kemp & Rotmans, 2005, p. 34). The term transition has gained prominence in comparison to system innovation and it has been argued that especially in the context of sustainability-oriented change and its proactive management, the term transition is to be preferred, because it is more processoriented. The notion of transition draws more attention to a desired end state and the long and complex change process leading there. The explicit focus on the process of change also directs attention to the concrete barriers for transitions and the various interlinked dynamics within and outside a socio-technical system that shape the process as a whole (Kemp & Rotmans, 2005, p. 36). So, transition and system innovation are used synonymously, yet they emphasize slightly different aspects or perspectives of the same kind of phenomenon. For the purpose of this thesis, both terms will also be used as synonyms, but the term system innovation is more central in the light of the research question and field of application. Since the potential of e-mobility developing as a sustainable system innovation or the question of whether a transition may be emerging from developments in the field of e- mobility depends on a specific core technology and the way it evolves with other nontechnical elements from the social environment in which it is embedded, the term system innovation seems more suitable. It takes the technology as a starting point and forces one to think about the long-term consequences of innovations: whether they give rise to or contribute to system innovation or do not alter the current path of development (Kemp & Rotmans, 60

74 2005, p. 35). Looking at e-mobility as a potential system innovation can thus help to distinguish in what ways the electric vehicle is not just another normal technological innovation and in how far e-mobility means a qualitative change in the deep structures of current mobility systems. Furthermore, the notion of system innovation facilitates a clearer focus on socio-technical co-evolution, rather than a purely macro-level focus on overall system levels, and it allows for focusing on the current state of an ongoing transition and analyzing its potential in the light of a wider and long-term future perspective. Apart from these basic notions that are the core or starting point of sustainability transitions as a distinct research field, other central concepts and propositions can be found in the different theoretical approaches and frameworks that are used to study transitions. Within the field of transition studies four to five different approaches of studying transitions have been identified. In their overview of the field, Markard et al. (2012, p. 955) identify four key frameworks in transition studies: the multi-level perspective on socio-technical transitions (MLP), technological innovation systems (TIS), transition management (TM), and strategic niche management (SNM). In their introduction to the journal Environmental Innovation and Societal Transitions and the field in general, van den Bergh et al. (2011, p. 9) identify four approaches as well, with a minor difference however: SNM is grouped together with the MLP and the fourth approach is suggested to be evolutionary economics with a focus on multiagent modeling. The multi-level perspective (MLP) on socio-technical transitions conceptualizes change in socio-technical systems as a process across three levels. The regime embodies the deep structure of a socio-technical system and contains its basic logic of functioning, while radical novelty emerges in niches that challenge the established regime structures. Regime and niches are embedded in a broader landscape level, i.e. the level of external influencing factors that shape overall developments. The MLP will be discussed in more detail in ch Strategic Niche Management (SNM) is closely related to the MLP and focuses specifically on processes within niches and conditions for their successful development. Since transitions in the MLP are seen as processes of change, where innovation and novelty originally emerge in niches, SNM focuses on these seeds of change and on ways of fostering their development. 61

75 Transition Management (TM) is closely related to the Dutch policy context as a term and a concept with a particular focus on theory as well as practice. TM focuses on the process of steering and guiding desired transitions towards a sustainable future. It is in this way similar to SNM, but goes beyond this approach by focusing on a cyclical process of managing transitions that concentrates not only on experimentation in niches, but also on creating shared visions and problem perceptions across a broad range of actors, and on up-scaling niche experiments and constant monitoring of the process as a whole. TM and SNM are discussed in more detail in chapter 3.2. The approach of Technological Innovation Systems (TIS) focuses on the relationship between technologies, or technological innovations, and the actors and institutions in a sociotechnical system, analyzing how these elements influence each other and co-evolve. Within transition studies, the focus is on contributions of specific technologies or technological innovation systems to sustainability transitions and the functions and processes characterizing successful technological innovation systems. The TIS approach will be discussed in more detailed in the following chapter Finally, multi-agent modeling approaches based on evolutionary economics are not featured prominently in the field of transition studies. However, it has been proposed that a more explicit engagement with evolutionary theory (which provides many basic concepts on which transition studies do build already) could add to conceptual and empirical approaches within transition studies as well as overall scientific rigor, and that modeling could be an important addition to the currently dominant focus on case studies and qualitative analysis (Markard et al., 2012, p. 964). For a more comprehensive overview of this particular stream currently emerging within transition studies, see van den Bergh et al. (2011, p. 11 f.) and Safarzynska et al. (2011). These different approaches of studying sustainability transitions have emerged from different disciplinary backgrounds, theoretical traditions and a plethora of combinations of different core concepts. These backgrounds and the key concepts, important for understanding the basic assumptions and premises guiding transition studies as well as those distinguishing the individual approaches presented above, will be introduced and discussed briefly in the following. 62

76 In general, transition studies with a focus on socio-technical systems have emerged from innovation studies, and especially those streams of innovation studies that have been influenced by evolutionary thinking that emerged from the crossovers between Darwinian biology and economics and that focus on industrial transformation and the development of technology, drawing also on the sociology of technology and, more recently, science and technology studies (Garud & Gehman, 2012, p. 981; Geerlings et al., 2009, p. 410; Stirling, 2011, p. 83). An early influence on transition studies is the work by scholars such as Nelson & Winter (1977, 1982) and Dosi (1982), who coined many of the terms and concepts dominating the field of transition studies today. As economists and innovation scholars they studied processes of technological development and the diffusion of innovations, and their particular scientific contribution was to introduce elements from evolutionary thinking into their field of study. For instance, technological development and the diffusion of innovations were explained in terms of variation and selection processes. Such basic elements of evolutionary thinking helped to make sense of typical phenomena of inertia and long phases of stability in processes of technological development. The concept of technological paradigm or regime was introduced, which denotes a structural element in shaping technological development, or, in evolutionary terms, the selection environment for new technological solutions or innovations. A technological regime has been defined to include the shared (e.g. among engineers or specific scientific communities) guiding principles, search heuristics, and cognitive routines, which shape processes and outcomes of technology development processes. It explains how some technological problems come to be recognized as such, how the options for solving them are selected, what is included in the potential solution space at all, and how the problem-solving process is being structured and carried out. The major achievement by scholars such as Dosi or Nelson and Winter is that they managed to show that these processes can (and need to) be explained if one wants to understand processes of technological change, because they are not given or determined by technology itself they rather emerge or evolve from specific cultures and rules among those developing technology. A technological regime can thus change over time, the state of knowledge and skills as well as cultural predispositions and perceptions of rational problem-solving approaches evolve as well (Garud & Gehman, 2012, p. 981; Geerlings et al., 2009, p. 406 f.). However, it can be shown that, due to these technological regimes exerting selection pressure, technological development is relatively stable over long periods. This is expressed in terms of path-dependency (Arthur, 1989; Basalla, 1988; David, 1985; Sydow et al., 2009) 63

77 and development following technological trajectories (Dosi, 1982). Similar to biological processes of mutation and variation, real novelty and radical innovations will from time to time, and basically by chance, originate as hopeful monstrosities (Mokyr, 1990, p. 291). In most cases, these will not manage to diffuse widely, because selection pressure exerted by established regimes will lead to their eventual disappearance. Nonetheless, sometimes there is fundamental technological change, usually triggered by external shocks or because a particular technological innovation manages to break through, either building on existing knowledge and skill or by destroying existing competencies and replacing them by new knowledge and skills. Any of these events can be the beginning of an era of ferment (Anderson & Tushman, 1990, p. 606) where the so far stable dynamics are disrupted, where numerous innovations emerge and compete for being selected and different possible future trajectories begin to take shape. In the ensuing variation and selection process, one technological option will emerge as the winner, and around this winner will emerge the next dominant design (Abernathy & Utterback, 1978; Arthur, 1989). Another relatively stable trajectory will form and the era of ferment is followed by a phase of incremental innovation, where the new dominant design is further stabilized by, for instance, improving efficiency and reliability of a specific technology, by working towards economies of scale and capturing large market segments (Garud & Gehman, 2012, p. 981; Sierzchula et al., 2012, p. 52). Such basic evolutionary conceptualizations of stability and disruptive change in technological development can still be found in more recent approaches in innovation studies. For instance, in his seminal paper on carbon lock-in, which is a central reference in transition studies, Unruh (2000, p. 820) builds on the concept of a dominant design and its evolution, in order to explain the emergence of specific technical systems, e.g. related to transport or energy supply, which cause today s severe environmental problems. An important insight is made here, which results from a critical perspective on evolutionary thinking: it is not automatically the best or (if possible to define this in any objective sense) superior technology that emerges as the winner from variation and selection processes and becomes the core of a new dominant design. Rather, it is a combination of technological characteristics, timing and framework conditions as well as strategic behavior that leads to a lock-in of new technological trajectories (David, 1985). This is an important insight with regard to the broadness of perspective on technological development and has consequences with regard to the role of involved actors and power constellations as well as the general economic and political framework conditions. It also shifts the focus when looking at concrete processes of transitions and disruptive technological change. For instance, another basic 64

78 concept in transition studies is the S-curve model developed in innovation studies (Rogers, 1962), which captures change processes along different phases, building on a predevelopment phase where niche actors, or in innovation studies terms early adopters, begin to develop innovative practices, followed by increasingly steep take-off and acceleration phases, evolving into a phase of stabilization where a new system configuration has emerged and the curve indicating dynamics and change flattens out again. While traditional economic approaches tend to focus on the second half of this curve and the long-run equilibria, insights from evolutionary thinking in innovation studies have shown that the earlier phases are decisive, in order to understand dynamics in an era of ferment and a situation where old trajectories are disrupted and a new dominant design begins to emerge through a complex variation and selection process (Hoogma et al., 2002, p. 198; Nilsson et al., 2012, p. 52 f.; Unruh, 2000, p. 820). Evolutionary thinking has not only contributed to the conceptualization of overall change processes in technology and innovation studies, it has also been used to develop a perspective on the many, not only technological, and interlinked drivers of innovation journeys. The concept of technological paradigms or regimes has already pointed to the interlinkages of technological innovations and institutions, i.e. guiding principles and cognitive routines, shaping technological trajectories. While in the beginning these regimes were mostly understood to be relevant within engineering or scientific communities, increasingly broader conceptualizations of technological development have included the overall societal context in which these processes are embedded. Thus, specific socio-technical configurations emerge from co-evolutionary processes including technical knowledge and technological innovation as well as political and economic circumstances, societal norms and values, individual behavior and use patterns (Stirling, 2011, p. 83; Unruh, 2000, p. 817 f.). Such a broader, co-evolutionary, understanding can for instance be seen in conceptualizations of long-wave theory on techno-economic paradigm shifts (Freeman & Perez, 1988). While in the beginning such shifts were explained as paradigmatic change in entire economies and societal institutions caused uni-directionally by technological innovation and economic forces, later work focused on mutual interrelations between different domains. Dynamic interdependencies between the societal dimensions of science, technology, economy, politics and culture were traced, in order to explain their alignment and inherent tensions causing shifts in techno-economic paradigms (Geels, 2011, p. 26). In transition studies, especially the multi-level perspective (MLP) on socio-technical transitions, and the related approach of strategic niche management (SNM), draw heavily on 65

79 basic concepts from this evolutionary perspective in innovation studies. The basic concepts of socio-technical regime and the derived niche and landscape levels have clearly been developed based on earlier versions of technological regimes or paradigms. Also, the MLP assumes path dependency and locked-in trajectories and conceptualizes transitions (to sustainability) as disruptive change at least in early versions of the MLP. As in evolutionary innovation theory, change is initiated by external shocks (in the MLP: developments at the landscape level) or radical innovations (in the MLP: at the niche level). Alternating phases of instability and change, as described by the notion of an era of ferment, and more stable phases where socio-technical regimes persist and innovations remain mostly incremental are also envisioned in the MLP (Garud & Gehman, 2012, p. 981). Co-evolutionary dynamics also play a central role in the MLP, first, across the three levels of niche, regime and landscape, and second, between different societal domains, similar to those identified in the literature on techno-economic paradigms, i.e. technological, economic, political, cultural, and scientific sub-systems (Grin et al., 2010, p. 4). A major difference when comparing the MLP and shifts in techno-economic paradigms is that the latter focuses on economies as a whole, while the MLP limits its scope to transport, energy, food or water systems (Geels, 2011, p. 26; van den Bergh et al., 2011, p. 8). A focus on socio-technical systems defined by the specific function they fulfill is characteristic of transition studies and is a way of dealing with the specific challenges of sustainable development, which go beyond individual technologies but can only be shaped in meaningful ways at a level below macro-level processes affecting entire economies and societies. The MLP has further been refined by integrating concepts from the sociology of technology and sociological literatures, referring especially to the work of Anthony Giddens. Based on these literatures, a distinction is made in MLP studies between socio-technical systems, containing the material elements such as technological artifacts and infrastructures, actors and actor networks, and regimes in terms of rule systems that actors refer to (Geels & Kemp, 2007, p. 442). This will be elaborated further in chapter While the approach of technological innovation systems has emerged from the same as well as newer, parallel streams within evolutionary innovation studies as the MLP and SNM, the Transition Management (TM) approach in addition draws on complex systems theory and concepts of resilience and adaptation. Based on an understanding of persistent problems precluding more sustainable development and a motivation to actively steer transitions towards sustainability, TM focuses on complex societal systems, governance approaches and principles of integrated assessment (de Haan & Rotmans, 2011, p. 90). Since concepts from these theories are combined with evolutionary approaches from innovation studies there are 66

80 sometimes inconsistencies in the systems perspective in TM (van den Bergh et al., 2011, p. 10). The notion of co-evolution is also referred to in TM, in the sense that managing transitions is a process where experimenting and learning are co-evolving elements in a cyclical process (Grin et al., 2010, p. 4). In short, in TM a comprehensive systems perspective is combined with reflexive governance approaches and principles of action research, resulting in a practice-oriented model for fostering and managing transition processes (Markard et al., 2012, p. 958 f.). Shortly summarizing this section, it can be said that apart from slightly deviating emphases e.g. in transition management in transition studies a socio-technical system is delineated according to the generic societal function it fulfills, e.g. mobility, energy provision or food production. A socio-technical system is further defined as a configuration of elements that include technology, policy, markets, consumer practices, infrastructure, cultural meaning and scientific knowledge (Geels & Kemp, 2012, p. 49). In this context, a system innovation refers to a long-term and co-evolutionary process of change, where all of these elements mutually influence each other, amounting to major shifts in the overall system configuration, i.e. a transition. Since this process includes not only physical changes in infrastructures or organizations, but also a redefinition of norms and values, new perspectives on how certain problems are framed and eventually acted upon, a transition can be defined as a substantial shift in the deep and underlying structure of a system (Elzen & Wieczorek, 2005, p. 651; Geels & Kemp, 2012, p. 49; Grin et al., 2010, p. 2 f.; Kemp & Loorbach, 2006, p. 105; Meadowcroft, 2009, p. 324; Rotmans & Loorbach, 2010, p. 108 f.; Shove & Walker, 2007, p. 2) The Role of Technology and Innovation Since the role of a specific technological innovation, the electric vehicle, is central in this thesis, some explicit attention is given to the role of technology and innovation in transition studies and the theories on which this field of study is built. The preceding section has provided a brief overview of evolutionary thinking in innovation studies during the mid- to late 20th century and many scholars of that time have again looked back to older theories of technological innovation and economic development. The most prominent sources for them have been the work of Kondratieff and Schumpeter. Kondratieff (1926) studied the interlinkages between technological and economic development and discovered that in the long term waves of development could be identified, spanning roughly 50 years. At the beginning of such a wave a groundbreaking technological 67

81 innovation can be found and when such a wave begins to decline, new inventions are stimulated and will eventually trigger the next long wave. Schumpeter (1939) built on Kondratieff s work and further elaborated the interlinkages between technological innovation and economic development. He explicitly focused on the link between a technological innovation and its impact on the beginning of a new period of economic upturn. The role of entrepreneurs in this process is emphasized by Schumpeter, because even though firms are willing to invest and take risks during an economic crisis, it takes proactive entrepreneurs who in a process of creative destruction turn the inventions of scientists and engineers into marketable innovations and foster their diffusion, which, in the case of radical innovations, can be the beginning of a new long economic wave (Dijk & Yarime, 2010, p. 1371; Geerlings et al., 2009, p. 404 f.). Building on Schumpeter s differentiation between an invention and an innovation, the role of specific types of technological innovations and development and diffusion processes have further been discussed in the field of innovation studies. As has already been mentioned, it is not always an obviously superior technology that successfully comes out of a diffusion process as the new dominant design. Radical innovations are thus not always the technologically most advanced or most sophisticated version of a specific technological artifact or machine, the more relevant question is whether a technological innovation manages to break through in a complex process of interlinking dynamics in a specific social context (Unruh, 2000, p. 820). Innovations can further be differentiated based on whether they build on and emerge from established trajectories and know-how, becoming disruptive only in the course of their further development, or whether they can be characterized as radical from the beginning, because they are making existing competences obsolete and require new rules and routines (Sierzchula et al., 2012, p. 52). For instance, Freeman & Perez (1988) distinguish incremental and radical innovations as well as changes of technology systems and changes in techno-economic paradigms. This systematization represents a distinction by scope of a specific innovation s impact: while incremental innovations are a minor technological change, radical innovations can have an impact on firms and industry structures. System innovations go beyond this level and affect also other societal domains, such as politics, culture and science. Finally, techno-economic paradigms shape entire economies or societies (Geels & Schot, 2007, p. 402). The approach of technological innovation systems (TIS) is making up a large part of transition studies and is the one approach that is most closely related to the shared origins of transition research in innovation studies. It focuses on the emergence of technological 68

82 innovations and specifically studies processes of socio-technical co-evolution, e.g. the interlinkages between technological development and related processes of institutional and organizational change. Innovation systems have been studied at different levels, i.e. sectoral, national, regional or technology-specific innovation systems. The notion of technological innovation systems (TIS), which is dominant in transition studies, goes back to Carlsson & Stankiewicz (1991). National innovation systems have been studied by some of the evolutionary innovation scholars mentioned earlier as basic references of transition studies as a whole, such as Dosi et al. (1988). A specific focus on sectoral innovation systems has been propagated by Malerba (2002), which aims at understanding sector-specific framework conditions, such as actor networks, institutions and knowledge base and how they shape technological innovation processes. Similar to other closely related approaches, such as social construction of technology (Bijker et al., 1987), or concepts such as large technical systems (Hughes, 1987), the basic assumption is that any socio-technical system, independent of its specific scale and scope, is characterized by technological and social or institutional elements that are mutually interdependent (Dolata, 2011, p. 19; Markard et al., 2012, p. 959; van den Bergh et al., 2011, p. 9 f.). In the context of sustainability transitions research, TIS studies are focusing on the emergence of radical innovations that contribute to sustainability-oriented change in sociotechnical systems, and consequently development processes of green technologies are researched in this field. This is a specific deviation from more traditional innovation system research, which tends to focus more on generic technologies, which may contribute to overall economic growth and development of national (or regional) economies. Conceptually, the focus of TIS studies has shifted from basic structures of an innovation system to specific functions that explain successful performance or failure of a specific innovation system, respectively. Based on case studies and historical event analyses, it has been studied what kinds of functions are important in technological change processes, where they worked well and how different functions are interrelated (Kern, 2012, p. 300; Markard et al., 2012, p. 959; Markard & Truffer, 2008, p. 597; van den Bergh et al., 2011, p. 9 f.). Usually, the implicit or explicit goal of TIS studies is to arrive at policy recommendations for supporting specific technologies, in transition studies recommendations are derived for supporting specific environmental innovations or green technologies. Often drivers and barriers for innovation, related to how well a specific TIS performs the relevant functions, are identified and based on this analysis, and sometimes in comparison with other TIS, recommendations are developed on how to improve performance or remove barriers for 69

83 innovation. Thus, the TIS approach is in most cases used to inform technology-specific policies (Markard et al., 2012, p. 959; Markard & Truffer, 2008, p. 601). However, it has also been discussed critically whether TIS approaches can really provide meaningful policy advice, when aiming at sustainability transitions. The focus on functions of innovation systems runs the risk of unduly neglecting factors and dynamics in the wider societal context. Similarly, the focus on one specific technology in a TIS study may lead to ignorance with regard to other technological developments that have an impact on the TIS under study, or maybe even compete with it. Finally, more emphasis should be put on explaining exactly why specific functions are not fulfilled in a TIS. In order to deal with these shortcomings of the TIS approach, it has for instance been recommended to find ways of integrating TIS and MLP perspectives (Kern, 2012, p. 300; Markard & Truffer, 2008, p. 610). To sum up this section on the role of innovations in transition studies, it has been shown that a focus on technological innovations and their development is dominant in TIS approaches, but also a basic element in other transition approaches. Radical innovations play a role because they may trigger changes in other system elements. However, the focus is on system innovations that may arise from such processes of socio-technical change. The scope of an innovation is also an important criterion. Since transition studies focus on sociotechnical systems, such as energy or transport systems, the scope is much smaller than for instance that of techno-economic paradigms, which shape entire economies, but broader than that of typical radical innovations in business and innovation studies, where the focus is on firms or industrial sectors. It is important to distinguish between different types of innovation, in order to avoid a perspective on transitions as deterministic and overly progress-optimistic technology-oriented projects. For the purpose of this thesis, a focus on TIS approaches has been discarded, because the electric vehicle is a good example for a radical technology that does not necessarily lead to a sustainability transitions. Since this is already obvious, it is not particularly useful to focus on conditions and functions of innovation systems for e-mobility (even if they included a focus on social innovation), because a broader conceptualization is needed The Multi-Level Perspective on Sustainability Transitions (MLP) A central theoretical approach in sustainability transitions research is the Multi-Level Perspective (MLP), which conceptualizes transitions in socio-technical systems as a dynamic interplay of processes across three levels landscape, regime, and niches that interact and reinforce each other (see Fig. 8). These do not refer to specific spatial or organizational 70

84 locations, but rather to a more theoretical idea of levels within functional space embodying different relationships between the respective actors, different dynamics and structures. The MLP framework draws on concepts from co-evolutionary innovation studies, science and technology studies, neo-institutional theory and social theory, in particular Giddens theory of structuration (1984). It allows for identifying characteristic patterns and mechanisms playing out in transition processes (Geels, 2011, p. 26; Geels & Kemp, 2012, p. 74; Grin, 2010, p. 4; Rotmans & Loorbach, 2010, p. 131). Fig. 8: Multi-level perspective on transitions (based on Geels & Schot, 2010, p. 25) In the MLP framework, the landscape provides a relatively stable environment characterized by large-scale developments and long-term trends, which can hardly be influenced by individuals or specific groups of actors. A regime is defined as a set of structure, culture and practices that is shared by a specific group of actors. The regime guides the behavior and actions of its members by shaping their perceptions of problems as well as the range of possible solutions. Overall, it conceptualizes a dynamic social structure that is firmly established because it is constantly reproduced yet, it also leaves room for a limited degree of variance. For new rules and routines to become part of a regime, individual and social learning processes are essential and issues of (normative) power need to be considered 71

85 (Shove & Walker, 2007, p. 4). Niches can develop where small groups of actors engage in new practices and behave in ways that do not conform to the general regime (Geels & Schot, 2010, p. 22; Kemp & Loorbach, 2006, p. 108; Nykvist & Whitmarsh, 2008, p. 1374; Rotmans & Loorbach, 2010, p. 108, 131 f.; Rotmans et al., 2001, p. 19 f.). Based on the idea that fundamental change has shaped the past and the present and that past experiences provide valuable lessons (de Haan & Rotmans, 2011, p. 90), the MLP has been applied to the study of historical cases, gaining a better understanding of fundamental change processes in retrospect, for instance in transport, shipping, aviation, water management and industrial production (e.g. Geels, 2002, 2005b, 2006a, 2006b). The second major field of application is the study of sustainability transitions currently underway or in the future, especially in the fields of energy, mobility and agri-food (cf. Geels, 2011, p. 29). Apart from that, there have been numerous research articles dealing with the conceptual and theoretical development of the MLP and the regime notion more specifically (e.g. Geels, 2010, 2011; Genus & Coles, 2008; Holtz et al., 2008; Markard & Truffer, 2008). Basic concepts and notions of transitions in the MLP framework will be discussed in more detail in the following The Concept of Regime Structuration and the Potential for Change The concept of regime is central to the MLP framework and will therefore be elaborated in some more detail. The regime might be seen as the core concept of the MLP because a transition (or a system innovation) is defined as regime change. Apart from this, the regime is the conceptual core of the MLP, in so far as landscape and niche may be viewed as derived concepts that are defined in relation to a specific regime (Geels, 2011, p. 26; van den Bergh et al., 2011, p. 10). As already mentioned, the earliest origins of the regime concept can be found in evolutionary economics and innovation studies, where the notion of technological regime was used to explain shared beliefs and common design heuristics among engineers and the emergence of technological trajectories. These approaches were further developed in the field of history and sociology of technology, studying how these technological developments are socially embedded, in so far as users, policy makers, existing infrastructures and institutions also shape paths of technological development (Dijk & Yarime, 2010, p. 1372; Geels & Schot, 2007, p. 399 f.; Genus & Coles, 2008, p. 4; Markard et al., 2012, p. 957). The concept of technological regime was then picked up within innovation studies and Science and Technology Studies, where the embeddedness of technology in society was further 72

86 emphasized. Schot (1998), Rip and Kemp (1998) refined the understanding of regimes and further developed the regime notion in the context of socio-technical systems. Geels (2002, 2004b) thus introduces the notion of socio-technical regimes and further develops its sociological basis, especially its foundations with regard to Giddens (1984) structuration theory. Regimes are conceptualized as rule sets that guide technology development at the micro-level, which is embedded in a context of macro- and meso-level developments of markets, infrastructures, regulations and culture. Introducing these different levels forms the basis of the MLP approach in current transition studies (Geels, 2010, p. 504; Geels & Kemp, 2012, p. 54 f.; Genus & Coles, 2008, p. 4; Smith et al., 2010, p. 436). Overall, it can be shown that across transition studies there is a common understanding with regard to the basic logic of the regime concept and the defining characteristics of a regime. In early conceptualizations of technological regimes in engineering communities as well as in the notion of regimes within socio-technical systems, the term regime is a helpful category for grasping stability and path-dependency explained as an inherent logic or direction underlying development paths. What is also similar throughout the development of the regime concept and throughout the transitions literature is a constant effort to better understand the sources for persistence and inertia in order to determine factors that explain regime change. However, there is also a major fault line separating two strands within transition literature, based on two different regime conceptualizations 3. They differ with regard to a more concrete understanding of what a regime actually is. First, there is a narrow conceptualization that defines regimes as rule sets. Second, there is a broad conceptualization that includes not only rules but also other elements, such as actors, infrastructures and technological artifacts, in the definition of a regime (Markard & Truffer, 2008, p. 605; Markard et al., 2012, p. 957). According to a broad conceptualization, a regime consists of actors and social networks, rules and institutions, technological artifacts and infrastructures. All of these regime elements are interrelated and emerge as a relatively stable semi-coherent configuration. Stability and path-dependency are created because the various processes within such a configuration of regime elements are interlinked and reinforce each other (Holtz et al., 2008, p. 626 ff.; Raven, 2007, p. 2198; Smith et al., 2005, p. 1493; Smith et al., 2010, p. 441; van den Bergh et al., 2011, p. 10; Verbong & Geels, 2007, p. 1026). 3 A similar argument has been developed in a co-authored conference paper: Best, Prantner & Augenstein (2012). 73

87 According to a narrow conceptualization, a regime consists of rule sets. A distinction can then be made between regulative, normative and cognitive rules. Different social groups (e.g. in the spheres of science, technology, politics, markets) have different rule sets and these rules guide the behavior of actors who reproduce the elements of a socio-technical system. In a regime, various semi-coherent rule sets are linked together and mutually influence each other. In this way, a regime is understood as the deep structure that lies behind the stability and path-dependency of socio-technical systems (Geels, 2002, p. 1260; 2004b, p. 904 f.; 2006b, p. 1071; 2011, p. 27). The rationale for a broad approach is to provide an all-encompassing view on all the material and immaterial elements related to fulfilling societal functions in an appropriate way (Markard & Truffer, 2008, p. 607). However, one might also argue that especially with regard to this complexity, the interpretive strength of a narrow definition of regime lies in the fact that it allows for a more profound understanding of the deep structure that lies underneath the activities of actors who reproduce system elements thus, structuring complexity (Geels, 2011, p. 31). In the end, the two conceptualizations may not be completely irreconcilable, since in practice it is difficult to conceive a pattern of socio-technical practices reproducing without the presence of institutions, just as it is difficult to see how institutions can persist without their re-enactment through networks of actors engaged in material practices as was pointed out by Smith et al. (2010, p. 442). From an analytical point of view, however, the difference between the two conceptualizations does matter and the choice depends on the research question as well as basic theoretical assumptions. A major difference between a broad and narrow regime conceptualization is the differentiation between the regime and the socio-technical system. In the narrow conceptualization of regimes as rule sets they are clearly separated. According to this view, a socio-technical system consists of material and measurable elements, thus, obviously, technologies and infrastructures, but also for instance regulations, consumption patterns and market relations. The regime as deep structure is then a more abstract category entailing the underlying rules that make sense of the basic logic of functioning of a specific socio-technical system (Geels, 2011, p. 31; Geels & Kemp, 2012, p. 56 f.). What both regime conceptualizations have in common, however, is their relation to a generic societal function as boundary criterion. This refers to very basic functions that are socially valued, such as energy provision, mobility or housing. The boundaries of a sociotechnical system or a regime, respectively, are drawn according to such a specific function. All the elements that are involved in the complex process of fulfilling such a particular 74

88 function are included (Markard & Truffer, 2008, p. 607; Holtz et al., 2008, p. 626 ff.; Smith et al., 2005, p. 1493; Smith et al., 2010, p. 436 ff.). Especially when referring to the narrow regime concept, it is important to note that the regime concept is based on Giddens theory of structuration (Giddens, 1984). Therefore the rule sets making up a specific regime represent Giddens notion of the duality of structure, in so far as these rule sets are drawn upon by actors in their decisions and actions (thus, shaping action), while at the same time actors reproduce those rules by way of their actions (thus, being an outcome of action). The basic contribution of Giddens approach is this very idea of duality, which is possible because structure and agency, or social practices, are sharply separated analytical concepts. Specific patterns of social practices are not themselves structure, structure is the more abstract principle behind concrete human action it only has a virtual existence (Giddens, 1984, p. 304). However, Giddens has been criticized for a lack of clarity and robustness when it comes to spelling out these abstract theoretical ideas in practice. For instance, structure is defined to consist of rules and resources, and when thinking about formalized rules or resources that provide agents with actual power over things or people, it can indeed be tricky to uphold the virtual nature of structure. An attempt has been made by Sewell (1992), arguing that structure as a virtual unit can be defended or reformulated by explaining the virtual interrelatedness between rules and resources: Schemas [i.e. rules] not empowered or regenerated by resources would eventually be abandoned and forgotten, just as resources without cultural schemas to direct their use would eventually dissipate and decay. Sets of schemas and resources may properly be said to constitute structures only when they mutually imply and sustain each other over time (Sewell, 1992, p. 13). Thus, the virtual nature of structure lies in the abstract principles and interrelations that help making sense of everyday objects and action. This type of ordering and clearly differentiating between tangible and intangible elements is similar to the separation in the MLP of a regime and a corresponding socio-technical system: the separation is analytical in nature and not valid in practice. Such an approach is useful when studying transitions because it shifts the analytical focus to underlying principles and explanations, however without resorting to structuralistic and deterministic models, but rather with a view to deeper meaning and explaining developments at a level that does not take observations at face value and forces the analyst to reflect on what is being observed and how one makes sense of it. Furthermore, it highlights the enabling aspect of structure, which is central when interested in the role of agency and questions of shaping future transitions. In essence, the duality of structure explains how, on the one hand, a system is constantly reproduced by actors, while on 75

89 the other hand, there is room for change because actors can reflect on the rules they draw upon. The three levels in the MLP are characterized by different degrees of such structuration (Geels & Schot, 2010, p. 18 f.). The landscape is characterized by the highest degree of structuration, i.e. absolute stability, because it represents a given environment, without a scope of action for individual actors aiming at changing landscape factors. Niches are characterized by the lowest degree of structuration in the MLP, they are highly unstable and they are the locus where truly new and different ideas that deviate from broadly shared frames of reference are being tested. Thus, niche actors actively create a space where their scope of action is as large as possible (Geels & Kemp, 2012, p. 52 f.). The regime is characterized by a relatively high degree of structuration: there is always some dynamic because actors can reflect on guiding rules and will introduce incremental changes in order to preserve and improve a socio-technical system; however, the basic rules are usually reproduced and attempts of developing a socio-technical system further generally remain within regime boundaries. Thus, the overall stability of a particular socio-technical system can be explained in terms of an existing deep structure or a regime. The landscape and niches can then be interpreted as derived concepts, which are defined in relation to a specific regime via their function as an external environment to it and as elements substantially different from it, respectively (Geels, 2011, p. 26 f.). Even though the regime concept is not applied coherently across different fields of transition studies, some general findings have been made. Various studies of either historical transition processes or of current transition processes have shown that established sociotechnical systems are relatively stable over long periods because they have emerged as specific configurations of technological, infrastructural and institutional elements that foster the development of incremental innovations rather than profound change that questions the underlying regime of a system (Markard & Truffer, 2008, p. 605; van den Bergh et al., 2011). So, how can change be explained in the MLP and what are basic theoretical assumptions regarding the potential for change in general? It has been shown that transition studies as a whole and also the MLP as one concrete approach draw on various theoretical traditions with different implicit assumptions as regards the potential for fundamental change. The MLP has, for instance, been criticized for being overly structuralistic with too little room for assessing the role and impact of actors and concrete agency in actual processes of change. This can be explained by the fact that the MLP is, amongst others, based on concepts such as techno-economic paradigms, which focuses on technological development of entire economies and studies such processes in terms of macro-level structural change. Furthermore, 76

90 socio-technical systems are defined according to the specific function they fulfill, which implies parallels with theories of structural functionalism (e.g. Parsons, 1951 or Luhmann, 1984), which have been discussed critically as regards their ability to explain phenomena of societal change (de Haan & Rotmans, 2011, p. 92). Simultaneously, however, the MLP is also firmly based on Giddens theory of structuration, and while this is indeed also one of the more structuralist approaches within sociology (Loyal & Barnes, 2011), building among others on Parson s work, its core focus is on the interrelations between structural elements and agency in concrete social practices. Giddens departs from Parson, arguing that there is no doubt that in his [Parson s] theoretical scheme the object (society) predominates over the subject (the knowledgeable human agent) (Giddens, 1984, p. XX). Giddens idea of the duality of structure and agency is emphasized here as a central tenet in the MLP and can thus help to distinguish it from more deterministic, structural perspectives on transitions as macro-level processes. It is used as an analytical perspective on currently on-going, highly uncertain processes and dynamics, as in the case of developing e-mobility, that helps understanding the role of involved actors, power constellations and concrete strategies. This is important for the question addressed here, namely understanding what the potential of current dynamics is, playing out in bounded time and space, to contribute to larger societal change processes. A closer look at Giddens theory of structuration and what it implies for questions of social change is needed, because at first glance the problem with the notion of structure is that it makes dealing with change awkward (Sewell, 1992, p. 2). The theory of structuration, similar to other grand social theories, e.g. that of Bordieu (1977) for instance, originally have been developed in order to explain the stability of a specific social order and the way that institutions are reproduced and societies persist. Phenomena of major change or transformation are thus usually explained as something exceptional and their causes are located outside the normal (Sewell, 1992, p. 3). That is also what the MLP is sometimes criticized for, being overly structuralistic, relying on landscape pressure, or disruptive niche innovations that fall outside the scope of what is normal, to explain change. This observation is accurate when taking the MLP for its face value, based on a superficial glance at the basic graph (see Fig. 8) illustrating multi-level dynamics. However, it will be shown that this is not the case when considering the theoretical foundations of the MLP rooted in Giddens structuration theory more clearly. Even though Giddens has not very explicitly dealt with phenomena of change or the conditions and processes of societal transformation, structuration also does not deny change altogether or is seen as a phenomenon of stability and persistence of concrete social 77

91 conditions. It captures the relation between agency and structure (which is persistent as an abstract process, not with regard to its consequences in practice) in everyday social practices, while these practices can result in a reproduction of current circumstances, or in a transformation of them just as well. Social reproduction must not be equated with the consolidation of social cohesion (Giddens, 1984, p. 24), which means that the ordinary operations of structures can generate transformations (Sewell, 1992, p. 16). Based on a critique of Giddens, Sewell argues that this is possible, if it can be assumed that structures exist at multiple societal levels; that actors can creatively apply rules and resources in different, sometimes unknown circumstances; that due to such creative application, the consequences for shifts in the allocation of resources, and thus power, is often unpredictable; that there are always multiple interpretations of the meaning of certain rules and resources, thus leaving leeway for creative actors to actively shape concrete social circumstances; and finally, if it can be assumed that different types of structures at different levels overlap or intersect, thus creating tensions and dynamic (Sewell, 1992, p. 16 ff.). Such a conceptualization is coherent with basic outlines in the MLP, where different levels are characterized by different degrees of structuration and tensions emerge between different sub-regimes or even between entirely different regimes (see ch ). It is also coherent with Giddens who assumes that social processes are in principle open-ended with regard to concrete outcomes. One explanation is the fact that actors are conceptualized as knowledgeable and purposive agents with the capacity to reflect on their motivations and action. At the same time, it is important to note that they are situated in a specific social context and that actors are not black boxes, but that their behavior can be explained by the interplay of structural conditions for action and the way that individuals deal with these conditions. Thus, structure is both constraining and enabling actors (Giddens, 1984, p. 26 f., 169; Schiller-Merkens, 2008, p. 192). Together with the fact that structural framework conditions may vary in time and space, that unintended effects occur, and considering the knowledgeability of actors who are able to consciously decide whether to conform or deviate from normal routines, change in societal systems is, in principle, possible and not necessarily dependent on external shocks or disruptions (Schiller-Merkens, 2008, p. 181). However, even though social change or transformation is possible through intended or unintended action by knowledgeable agents, it is not possible to identify specific generalizable causalities or factors that influence structuration processes and lead to societal change. Such general theoretical propositions are not feasible because structuration processes are contingent and only become meaningful in concrete social practices and situations. 78

92 Similar to basic understandings of transitions, Giddens conceptualizes social change as a multi-causal phenomenon that can only be identified in retrospect by identifying start- and end-points, by some type of context-specific indicators, in an ex post analysis. It is further argued that due to the decisive role played by knowledgeable agents, transformation processes can only be understood by studying not only changing social structures but particularly strategic behavior by involved actors and the way that they refer to and position themselves with regard to structural framework conditions (Schiller-Merkens, 2008, p. 183 ff.). Understood in this way, structure is not a deterministic concept explaining first and foremost stability and reproduction processes, and much less a teleological idea assuming simple causalities for explaining complex transformations rather, the concept of structuration can be particularly helpful in any critical analysis of the dialectical interactions through which humans shape their history (Sewell, 1992, p. 27). Since transition studies, and especially MLP analyses, are sometimes criticized for neglecting the role of concrete actors, this reconsideration of and explicit reference to Giddens is helpful and the case study on e- mobility will explicitly focus on strategic behavior by the involved actors Transition Patterns and Pathways More structural, basic dynamics of transition processes have been developed in MLP studies (while the role of specific actors is implicitly assumed as integral part of structuration processes). Most prominently, Geels and Schot (2007) aim at refining the MLP by introducing more differentiated transition patterns and pathways, in order to deal with weaknesses and simplifications the original or basic version of the MLP has been criticized for. A major criticism has been that the MLP embodies an overly simplistic understanding of change processes where niche innovations emerge and subsequently become dominant, eventually overthrowing an existing regime. This basic conceptualization has led to a large array of transition case studies, which mainly focus on specific, presumably green, innovations and the way that they struggle against established regimes, with regimes being treated as monolithic, undifferentiated barriers or prohibiting structural conditions. In order to overcome this so-called bottom-up bias and illustrate the complexity of transitions, the MLP has been enriched by focusing on transition patterns and characteristic pathways, emerging from varying dynamics in nature and time throughout transition processes (Geels & Schot, 2007, p. 405; Geels, 2011, p. 32). A basic differentiation of change processes has been introduced by Geels and Kemp (2007). They distinguish reproduction, transformation and transition processes. First, reproduction processes characterize stable regimes where incremental innovations contribute 79

93 to dynamic stability and regime preservation by actors continuously reproducing established regime rules. Second, transformation processes occur when there is regime change, enacted from within the regime but influenced by landscape pressure. Finally, transition processes occur when there are dynamics across all three levels leading to a new socio-technical system through landscape pressure, weakening of the established regime and the diffusion of radical innovations emerging at the niche level. Differentiating between these types of processes reflects some of the basic difficulties when dealing with structuration and change as discussed in the previous chapter. The three processes are differentiated according to structure-agency interrelations, i.e. the type of structural conditions and dynamics on the one hand, and the way that actors get engaged on the other hand. Furthermore, a balanced view on influential factors and causes for systemic change is developed, arguing that transition as well as transformation processes are predominantly emerging from inside a socio-technical system, while external developments need to be considered as important influencing factors as well. Thus, this approach illustrates a specific way of dealing with problems of structuralistic approaches, i.e. by stressing the interlinkages between actors and structure, and system-internal as well as external causes in processes of change (Geels & Kemp, 2007, p. 453 f.). A typology of transition pathways has been developed by Geels and Schot (2007), further differentiating variations of timing and nature of multi-level dynamics in transitions. Based on historical case studies, four basic transition pathways are identified. They show differences in timing of interactions between niche, regime and landscape levels, the nature of these interactions ranging from symbiotic to disrupting, and the combination of timing and nature of specific dynamics (Geels & Schot, 2007, p. 405 f.). Table 2 provides an overview of the four basic transition pathways. It should be noted that these pathways may overlap or emerge in combinations in actual cases in practice. These pathways are ideal-types, derived from basic theoretical propositions and illustrated with specific historical cases (see for instance Geels, 2002, 2005b, 2006a, 2006b, 2007). However, they will rarely occur in practice in such ideal-typical fashion, there may be variations and, with a view to ongoing transitions, it will not be possible to determine whether a full-blown transition will unfold or predict future dynamics. Therefore, the pathways can guide empirical analyses but have to be applied carefully, focusing on the different dynamics and mechanisms at play as well as the particular internal logic emerging in actual pathways (Geels & Schot, 2007, p. 415 f.). 80

94 Tbl. 2: Typology of transition pathways, based on: Geels & Schot, 2007, p. 406 ff. Apart from such broad transition pathways, more specific transition patterns have also been identified. First, an add-on and hybridization pattern can be observed where niche innovations are step-wise integrated in regime configurations and lead to the emergence of new combinations, e.g. the integration of battery technology in cars, ranging initially from starter batteries, to a combination of battery and internal combustion engine in hybrid cars, to pure battery-electric vehicles. Second, knock-on effects and innovation cascades an be observed where a technological innovation triggers change in related technological fields or leads to increasingly radical changes through interlinkages with other social phenomena, e.g. in the case of e-mobility, renewed interest was caused by developments in battery technology in non-automotive applications as well as its interrelatedness with energy transition issues and efforts towards more livable cities. Third, a fit-stretch pattern emerges where an innovation that was developed as a technological substitute able to fit in current system configurations over time turns out to be a radical innovation, e.g. a stretch may occur from developing electric vehicles to broader conceptualizations of e-mobility as part of a fundamentally different transport regime. Fourth, hype-disappointment cycles are common phenomena, especially in the field of alternative vehicles where in the beginning expectations are high, causing a hype around the new technology, often resulting in disappointment when it cannot meet all those expectations. This pattern illustrates for instance the alternating phases of interest that can be observed for electric vehicles and fuel-cell vehicles. Finally, a nicheaccumulation pattern can be observed where a niche innovation subsequently moves from its initial small technological niche (R&D or demonstration projects) to ever larger market 81

95 niches, which is for instance also envisaged for electric vehicles being developed and tested in different fields of applications across various market niches, ranging from pilot projects to fleet applications in specific market niches (Geels & Kemp, 2012, p. 61 ff.). A similar approach, focusing on the individual event sequences and generalizable, shorter-term mechanisms against the background of a broader transition dynamic, has been proposed by de Haan and Rotmans (2011) based on the assumption that any transition pathway consists of distinguishable patterns. This approach is explicitly distinguished from Geels and Schot (2007), because it is related to transitions thinking based on complex adaptive systems and classic systems theory, which is a source of transition management rather than the MLP. Furthermore, the focus is on individual patterns, rather than transitions as a whole. However, the basic approach of deducing basic propositions from theory is similar, and even though the approach is different from Geels and Schot, it is argued that basic transition thinking is similar and that since a transition pathway also is a transition tale it ought to be possible to retell it with the patterns proposed in this article (de Haan & Rotmans, 2011, p. 91). Three patterns are identified reconstellation, empowerment, adaptation which are basically distinguishing top-down, bottom-up or system-internal impulses for change. Such a systematization is clearly derived from a conceptualization of societal systems as complex adaptive systems. According to classical system theory, change is caused by factors outside the system and by dynamics that are not normal. Since the system in question is a societal system, external influencing factors can be characterized to work either from top-down or bottom-up. The conceptualization as a complex adaptive system allows for a third classification, i.e. change induced from within the system via e.g. feedback loops and gradual evolution as a normal process. So, first, a reconstellation pattern describes a top-down pressure for change, from outside a specific societal sub-system or by government. Second, an empowerment pattern describes the processes where niche-innovations rapidly break through and become viable alternatives for an existing regime. Third, an adaptation pattern occurs when the system is changed from within, in order to secure its persistence or optimize the way it functions. This may sometimes require quite fundamental re-structuration or innovation processes, in the case of adaptation they are initiated from within established regimes (de Haan & Rotmans, 2011, p. 96). Even though these patterns are different from the work of Geels and Schot (2007) with regard to conceptual background and analytical focus, they can indeed be found within the more elaborate transition pathways as well. Also, both approaches may seem to oversimplify 82

96 transition dynamics in the real world (especially the very basic patterns identified by de Haan & Rotmans (2011)) and may be criticized for simultaneously explaining everything and nothing, due to their very general and abstract nature. However, this type of conceptualization can make sense, when considering the basic aim and explanatory style of transition studies in general, and MLP analyses in particular. When studying long-term complex phenomena, such as transitions, a broad heuristic framework, such as the MLP, is suitable. However, in order to arrive at generalizable and valid results when studying such phenomena, heuristic and process-oriented approaches need to produce patterns or mechanisms, i.e. repeatedly and typically occurring, causally linked event sequences. This is also important for systematization and comparability across qualitative case studies or narratives and may pave an epistemological middle way (Geels, 2011, p. 35 f.) between quantitative and variable-oriented research on the one hand, and more constructivist methodologies on the other (de Haan & Rotmans, 2011, p. 96, 100; Geels, 2011, p. 35 f.) Niche-Regime Dynamics The discussion of transition patterns and pathways, which have been developed to refine the MLP and explain typical dynamics of change in more detail, shows that, in principle, the various types of interactions between niche-, regime-, and landscape-levels are decisive for understanding transitions. However, the concrete patterns and pathways also imply that interactions between niches and a regime are of particular relevance, because a transition is, in essence, a change of regime, and the specific type of change, the nature of the new regime depends on the specific kind of variation emerging in niches, from which new components can be selected. Niches thus provide the seeds for change; they determine the range of alternatives that have been nurtured in protected spaces, which are available for new trajectories to develop at all (Smith, 2007, p. 429). While landscape developments can be both disruptive or stabilizing, the interactions between niche and regime can be much more varied and determine in a more qualitative fashion what exactly changes and how. In transitions, a niche-innovation (which can be anything, ranging from a specific technology, to an alternative social practice, new rules or guiding principles, or even changing societal norms) has to replace an existing regime. It does so either by overthrowing it in a disruptive change process, or in a more step-wise incremental way, sometimes even actively enacted from within the regime, e.g. by slowly adjusting established rules, or optimizing a regime continuously or only in some places, which in the end accumulates and amounts to more fundamental change. Without a doubt, the landscape is important as well historically, 83

97 factors such as macroeconomic crises, environmental catastrophes or global cultural development trends have always been important for triggering societal change. However, if one wants to understand especially shorter-term phenomena and ongoing developments that might only someday turn out to have been part of a broader transition, niche-regime dynamics are of particular relevance for a number of reasons: First, a focus on niche-regime dynamics means a focus on concrete actors and their strategies. While the landscape level contains broader trends or disruptive crises that cannot be influenced directly by actors and which are emergent (social or ecological) phenomena, actual strategies and practices that shape concrete systems and room for actively guided change or persistence exists only at the niche and regime level because they describe degrees of structuration in the social practices carried out by actual people. The landscape which is a bit clumsily conceptualized as the level with the highest degree of structuration, thus containing no scope for direct human action at all, is in fact different in nature and indeed external to processes of social change: While niches and regimes work through sociological structuration, sociotechnical landscapes influence action differently. Sociotechnical landscapes do not determine, but provide deep-structural gradients of force that make some actions easier than others (Geels & Schot, 2007, p. 403). Thus, when interested in shaping transition dynamics or the socio-political processes involved, the more agency-prone categories of niche and regime are particularly relevant and helpful in guiding empirical analyses. Second, especially when focusing on (potential) transitions in the making, a focus on niche-regime dynamics is important. A transition can only be identified in retrospect and with landscape developments it is clear that these macro-level phenomena and their actual impact will only be observable in retrospect. This can be illustrated by a relatively recent example: Whether the nuclear catastrophe at Fukushima will have had a role to paly in a broader energy transition can only be determined in retrospect while the fact that the German government has politically decided that an energy transition is to be carried out is a much more relevant and observable phenomenon, which can be analyzed in its concrete form or choice of instruments, allowing for a discussion of the suitability of concrete policies and potential alternatives. Third, the example of the German energy transition and Fukushima also highlights another aspect in favor of focusing on niche-regime dynamics: landscape developments or concrete events, such as the Fukushima incident, have no independent meaning of their own. Whether they exert pressure on existing regimes or stabilize established development paths 84

98 depends on how they are perceived by concrete actors and what they make of it regarding their strategies and social practices. Landscape developments may come to be seen as something that confirms established strategies, thus contributing to regime stabilization and persistence, or they may make established strategies or guiding principles seem to have lost their effectiveness and raise questions about adapting behavior or thinking about radically new ways of doing things, thus contribute to the emergence of niches. Whether a landscape event or trend is seen as an indicator for a weakening regime and, what is more, as a concrete pressure for change, depends on perceptions and interpretations, e.g. in the form of policy formulation (Smith, 2007, p. 430). In the case of Fukushima it can be argued that even though it is used as legitimization for a government-induced energy transition project focused on renewable energies, it could in theory have just as well been used as an argument for increased efforts in nuclear R&D or a political focus on technologies improving safety and reducing risks. Finally, it should be noted that some caution is warranted when studying niche-regime dynamics in concrete empirical cases, especially with regard to the definition of what exactly constitutes a niche or a regime. When interested in actors and strategies in transition processes and applying a multi-level perspective, one tends to group actors as regime actors or niche actors. This is problematic because this is often done in transition studies, and it is done without explicitly laying out the reasons for assigning a niche - or regime -affiliation. However, this has consequences for the subsequent analysis and it also contains implicit assumptions regarding problems and research questions as well as judgments about barriers and drivers for change (Pel & Boons, 2010, p. 1251). Furthermore, it is problematic from a conceptual point of view to describe specific actors as niche- or regime-actors. Levels in the MLP describe degrees of structuration and thus the deep structure that can, as a metaphor, help to better understand the basic principles explaining specific routines, strategies or social practices. It is thus important to differentiate between concrete actors and a regime embodying established structures or deviating activities that can be characterized as niches. The example of e-mobility clearly shows that the delineation of niche and regime is not clearcut, neither in terms of actors nor in terms of structure. For instance, e-mobility can in practice be part of established regimes of car-dependency and private car ownership, or an element of sustainable urban mobility based on carsharing and intermodality, or the central technology functioning as the interface of a completely new system linking mobility and energy systems. Similarly, large car manufacturers are involved in developing electric vehicles and thus contributing to an emerging e-mobility niche, while they predominantly 85

99 reproduce regime structures in their day-to-day business. So, even though delineations are often overlapping or fuzzy, also in this thesis it will be referred to niche- or regime-actors, in order to express basic power relations or strategic orientation. Nonetheless, it should be noted that niches or regimes do not describe, in theory, concrete places or people. Of course, a strict separation of structure and agency in practice is not possible and only makes sense for analytical clarity. For instance, differentiating between different groups of actors and degrees of structuration at niche or regime levels may help avoid a bottom-up niche bias (discussed in ch ) where, in some cases unduly oversimplified, the focus is on green niche innovations struggling against regimes as monolithic barriers for change. It is therefore important to focus on the concrete and case-specific interactions between the levels of niche and regime. Some basic concepts and relevant processes have been identified in transition studies, focusing more in detail on the way that niches gain momentum and in what ways niche and regime dynamics are mutually interlinked instead of simply being opposed and struggling against each other. For instance, the concept of nicheregimes is used to describe a situation where a specific niche has grown into a visible alternative being in actual competition with regime structures, while not (yet) replacing an established regime as a whole. This conceptualization highlights the fact that there are intermediate dynamics or stages and that there are varying degrees of tension between niche and regime. A conceptualization of niche-regimes may also help to focus on the radical innovations and alternatives that emerge from smaller niches, which are clearly opposed to the established regime, and the viable competing options to established regimes instead of the more symbiotic innovations or smooth change processes that are part of normal regime adaptation and optimization processes (de Haan & Rotmans, 2011, p. 93). This juxtaposition of radical and more symbiotic niches also highlights the basic dilemma in niche-regime interrelations. The chances for a niche to diffuse widely are highest when there is some degree of fit with established regime structures. While in some cases a smooth and step-wise integration of niche-innovations into an established regime may over time and in the course of learning processes lead to regime change, it is more likely that the result is an incremental adaptation of the regime without the more fundamental kind of change characterizing a full-blown transition. In contrast, a more radical niche that questions and opposes basic regime structures will rarely manage to break through and diffuse widely, especially when the regime is stable and not, for instance, weakened by landscape pressure. This type of niche contains the potential for fundamental change envisioned in transitions, but it will often disappear, because it finds no way of linking itself to the regime and growing into 86

100 more dominant structures (Smith, 2007, p. 430). Zooming in on concrete niche-regime dynamics, variations and cases in between these two extremes or ideal-types can be identified. One approach of studying the concrete dynamics involved in niche-regime relations is to focus on processes of translation between niches and regimes and how these mutually influence developments in niches and regimes (Smith, 2007). Three basic forms of translation have been identified: First, second-order learning processes lead to the identification of (sustainability) problems that are rooted in the deep structure of a regime and this informs the process of niche-building. In this way, regime structures are translated into a different frame of reference with different problem perceptions that form the core of an emerging niche (Smith, 2007, p. 443, 446). Second, the radically different visions emerging in niches are in some cases translated back into regime language. The niche has shown that alternatives to an established regime are possible, or can at last be imagined, and the subsequent task then is to convey this to the regime level. Usually this means that elements within a niche are sought that are flexible enough to be interpreted in regime terms as well. For instance, if e-mobility is reduced to an issue of drive technology, it can simultaneously be part of a more alternative niche and the established regime. However, this example shows that [u]nder this kind of translation, there is wider diffusion of a more shallow sustainability (Smith, 2007, p. 446). Third, there are processes of mutual translation where regime incumbents and niche actors are more directly involved with each other, e.g. in large demonstration projects, real-world experiments or transition management projects. Ideally, in these settings processes of second-order learning are facilitated and a mutual understanding of the different basic assumptions and guiding principles can be reached. Eventually this may lead to a growing niche-regime, increasing the potential for a wider transition (ibid.). Like translations, expectations and future images can play an important role, especially when niche-regime dynamics concerning a specific technology are concerned, as in the case of e-mobility. Some lessons have been learned with regard to images of the future role of new technologies, which can be used to argue that a cautious analysis of ongoing dynamics between niche and regime is important. Similar to the translation process from the regime to the niche level, future images of technologies are in the beginning shaped by established regime structures: current cultural norms and values shape expectations and new technologies are envisioned as substitutions of existing technologies serving unchanging needs. The possibility of completely new social practices emerging around different needs in the future or a new combination of old and new technologies is often ignored. At the same time, niche actors typically engage in translation processes that focus on promoting the 87

101 specific technology, around which the niche is formed. Since they aim at fostering diffusion, they usually exaggerate the advantages of a new technology, and this may be successful for a while, especially when a new technology is suddenly appearing or experiencing development leaps and thus triggers fundamental shifts in visions of the future. When high expectations eventually cannot be met, which is likely for radical innovations that cannot easily be integrated in an established socio-technical system, disappointment and disillusionment can weaken the niche substantially (Alkemade & Suurs, 2012, p. 450; Budde et al., 2012, p. 1074). This basic discrepancy in future visions and in the related niche-regime dynamics, is also expressed by the fact that processes of diffusion and societal embedding of a new technology are usually not problematized to a sufficient degree (Geels & Smit, 2000, p. 877 ff.). Thus, images of the future may vary within and between niches and the regime, which may lead to an eventual weakening of the niche, or possibly also to these images serving as boundary objects of translations between niche and regime levels. This type of dynamic also implies that the boundaries between niche and regime can become blurred or shifting. This is highlighted by the example of large car manufacturers investing in R&D for e-mobility while at the same time there are visions of future e-mobility that is more sustainable also in terms of more intermodal transport or carsharing, which threatens the basic business models of the automotive industry. Even in such a case, which at first glance amounts to not much more than technological development and diffusion of electric vehicles, various socio-technical configurations are emerging and these cannot always clearly be assigned to the regime or a specific niche. It also stresses the point that specific actors cannot easily be classified as regime- or niche-actors, not only due to conceptual considerations, but also because a strict delineation of niche and regime is not always feasible. For empirical analyses it is thus important to focus on the way that specific actors position themselves, e.g. in translation processes, and on actors that move between niche and regime levels, e.g. termed pragmatic system builders (Smith, 2007, p. 447), facilitating social learning Multi-Regime Dynamics Apart from sometimes fuzzy boundaries between niche and regime in transition studies in general, the specific case of e-mobility furthermore cuts across multiple socio-technical systems and the associated regimes. Multi-regime interactions have been identified as an understudied, but promising field within transition studies, because a number of environmental innovations have emerged that cut across regime boundaries, such as the 88

102 electric vehicle but also biofuels, combined heat and power (CHP), or using biomass from waste for generating electricity (Geels, 2011, p. 32; Konrad et al., 2008, p. 1191; Nilsson et al., 2012, p. 57). Developing e-mobility as a new drive technology and form of personal mobility affects first and foremost established transport or mobility regimes. Apart from that, the energy system plays a central role as a newly emerging actor in the field of mobility, providing electricity for powering electric vehicles. Especially when considering the potential of electric vehicles providing storage capacity in the context of an energy transition based on renewable energy sources, the energy regime is itself affected by developments in the field of e-mobility. Integrating electric vehicles in energy infrastructures will most likely require new mobility and energy services that go beyond traditional business models in energy provision, thus affecting energy system regimes. Finally, information and communication technologies (ICT) will become more important with regard to linking vehicles and energy infrastructures as well as different modes of transport and facilitating mobility services, which is relevant where developing e-mobility includes intermodal transport and carsharing services. Apart from cases such as e-mobility, where a central technology creates new links and is located at the intersection of otherwise unrelated regimes, there are other cases where, for instance, two separated regimes are linked by a third regime (Konrad et al., 2008, p. 1197). One might also argue that e-mobility is such a case, in so far as ICT functions as an enabler of e-mobility by providing the possibility to link electric vehicles and energy infrastructures. Distinct regimes may also be identified, which fulfill a similar societal function and thus directly compete with each other. An example in the field of mobility is when a distinction is made between public and private transport regimes, or between separate regimes related to the different modes of transportation. Less competitive multi-regime dynamics may be observed where different regimes are exhibiting similar structures. In such a case there may be learning processes or synergy effects resulting from multi-regime dynamics (Konrad et al., 2008, p. 1198). In the case of e-mobility, it can be argued that transport and energy regimes are currently struggling with similar structural challenges. Both sectors are faced with increasing competitive pressure and concentration processes: in the energy sector due to deregulation and liberalization of energy markets and in the automotive sector due to saturated markets, at least in industrialized countries, and overcapacity. Developments in the field of e-mobility may thus either lead to symbiotic developments towards sector convergence, or to a situation where they enter into competition for market shares in a newly emerging e-mobility regime (Rammler, 2011, p. 19). 89

103 A systematic typology of multi-regime dynamics has been developed by Raven & Verbong (2007). They differentiate between four basic types of multi-regime interactions. First, there may be a competitive relation where two or more regimes fulfill a similar societal function. Second, there may be a symbiotic relation where regimes are closely connected and depend on each other or mutually benefit from each other in fulfilling their specific societal function. Third, there can be integration of formerly separated regimes, or parts of regimes, into a single new regime. Finally, there can be spillovers from one regime to another, e.g. when they are structurally similar, linking them together by learning and adaptation processes (Raven, 2007, p. 2199). The examples related to the case of e-mobility highlight that the regime concept can be applied to various empirical objects of study and levels of analysis always depending on the research question. For the purpose of this thesis, the role of the electric vehicle as a new technology is central and it is asked whether socio-technical change patterns evolving around this technology can lead to more sustainable forms of mobility. This includes a view on the energy system, thus the focus is on multi-regime dynamics emerging around a radical innovation connecting different regimes rather than an enabling regime (e.g. ICT) linking two otherwise unrelated regimes (e.g. energy and transport). Furthermore, the focus on sustainable mobility in a broader sense leads to a definition of an overall mobility regime, rather than a very detailed assessment of different transport sub-regimes (e.g. car-based, public, walking and cycling) and their basic rules of functioning. In this thesis, societal function is defined more generically, as providing mobility and the different transport modes are understood to be different elements of an overarching mobility regime, which of course includes various sub-regimes, not only related to different modes of transport, but also the political and regulatory framework conditions, market and industry dynamics, science and R&D, and user behavior in the field of mobility. A similar case where a specific technology has played a central role in emerging multi-regime dynamics is that of biomass from waste used as a source for electricity generation in the Netherlands. The formerly separated waste and electricity regimes began to interact when new ways of incinerating waste became interesting for energy companies looking for green energy technologies. A symbiotic relationship began to develop where waste companies also become electricity producers and energy companies also collect and process waste as a renewable energy source. A decisive aspect in this process was that the basic idea of waste has been reframed and the new concept of biomass has emerged as the binding element in the interaction process (Raven, 2007, p. 2206). Such a fundamental 90

104 change in what waste is understood to mean, turning it into a valuable resource in an emerging energy system based on renewable resources, also means that the involved actors, in this case waste processing companies and energy producers, re-defined their roles and reorganized part of their business models and strategies. Thus, basic frames of reference that in the past separated two different regimes co-evolved and changed into a new and more symbiotic relationship between those regimes (ibid.). It is proposed that a similar situation can be expected for the case of e-mobility, because the electric car as a technological artifact is a factual link between energy and transport regimes and, what is more, a broader understanding of e-mobility may evolve as a new concept, similar to that of biomass, that implies also deeper changes as regards role perceptions and basic functions. Again, there is a major difference between a situation where electric vehicles diffuse and substitute for conventional cars with the new fuel provided by energy companies rather than the oil industry or, in contrast, a situation where e-mobility emerges as a system innovation with fundamentally different mobility patterns, business models, technological infrastructures and socio-political implications. Why should multiregime dynamics be considered as a potentially important element in the emergence of system innovations (in the case of e-mobility, or also the presented biomass case)? Referring back to Giddens and its central tenets in the context of the MLP, it is argued here that the emergence of a technological artifact such as the electric vehicle that creates linkages between formerly separated regimes can have a potential for inducing second-order learning and structural change. In a setting where multi-regime dynamics unfold, actors are confronted with the diverging logic of another regime, its modes of signification and legitimation, its established patterns of resource allocation and accepted power constellations, and they are thus in a way forced to reflect on their own premises and guiding principles, in order to be able to make sense of the new situation (Schneidewind, 1998, p. 318; Schneidewind & Petersen, 1998). At least in principle there is a chance for opening up creative space, where tensions and processes of reframing basic rule systems and frames of reference lead to a greater potential for the emergence of a system innovation. If such potential came to be realized and a system innovation were to emerge, it is hardly possible to predict or assess ex ante what the new socio-technical system and regime would be like as regards its concrete shape and basic structure. This is another aspect that makes a focus on multi-regime dynamics essential for the case of e-mobility and, in fact, any future oriented analysis of potential system innovations. Looking at historical cases of transitions or system innovations it is possible to, ex post, identify a specific regime that has 91

105 emerged and trace the way it has developed, identify patterns of change that have played out over the course of history, and in this way also observe how boundaries of previously separated regimes may have become blurred or integrated into a new regime (as in the case of biomass, for instance). This is much less possible when one is to carry out an ex ante analysis of a potential future transition. Since it is not clear what a future regime may look like, and its emergence will result from a contingent process, one has to be careful to broaden the analysis beyond the regime in question (in this case the mobility regime) and also take account of other regimes that may have an impact on it or newly emerging interrelations with it (in this case the energy regime, for instance) (Konrad et al., 2008, p. 1191, 1197). In sum, e-mobility is a challenging but also promising case from a multi-level perspective. It highlights the difficulties in delineating niche and regime, since national governments and large car manufacturers are involved in developing the e-mobility niche. In addition, it is a case where multi-regime dynamics have an important role to play, due to the new linkages created. It is thus important to focus on e-mobility as a field of innovation that cannot meaningfully be analyzed by a traditional focus on sectors or a specific regime. Especially regarding its potential for a future system innovation, it is imperative to pay attention to the many interlinkages e-mobility creates between, first and foremost, energy and mobility regimes, and also the factors fostering or hampering the development of e-mobility, which may be found across different regimes Critical Reflection: Shortcomings and Limitations Criticism of two sorts should be discussed here: First, a critical perspective will be presented on the theoretical foundations of the MLP and its value in conceptualizing and explaining transformation processes. Second, criticism relating to applications of the MLP in transition studies as an emerging field of research will be discussed. These two different dimensions sometimes get mixed up and may in some cases lead to an unduly harsh critique or rejection of the MLP. In order to understand the basic theoretical value, and the limitations, of the MLP on the one hand, and to provide a critical perspective on the way that the MLP has been applied on the other hand, these two aspects should be treated separately. Even though a clear separation is not always possible, a distinction in principle is important, because the first type of criticism relates to conceptual difficulties and limitations, while the latter basically has to do with weak application. In sum, the previous chapters have shown that the MLP provides an overarching heuristic framework that has a particular strength in analyzing transition processes in socio- 92

106 technical systems defined by their specific societal function. The regime concept is thus a helpful analytical category in cases where the research question is concerned with future more sustainable alternatives to fulfill these functions and not primarily in the fate of a specific technological configuration (Konrad et al., 2008, p. 1192). Despite this broad aspiration, one weakness of the MLP approach is its view on co-evolutionary processes of change. In many transition studies it is analyzed how green niche innovations struggle against established regimes and the focus seems to be on the necessary changes at the level of regimes that would allow a transformation built around the specific niche innovation rather than a clearly focused analysis of co-evolution of technological and social elements across niches and regimes. In that sense, van den Bergh et al. have quite insightfully characterized the MLP approach regarding transitions as a co-evolutionary (or, more accurately, codynamic) interplay between processes functioning at the different levels of landscapes, regimes and niches (van den Bergh et al., 2011, p. 10). It can be argued that evolutionary approaches and innovation studies as dominant theoretical background have introduced a bias towards the role of technological innovation, i.e. a focus on niche innovations and the way that they get selected through evolutionary processes and overturn a regime. This critique is not to say that technology as a factor is to be ignored, which would be a wrong conclusion. Technology can be very powerful indeed, but it is not independent of its social surroundings. It is important to understand the interrelations and co-evolutionary dynamics, because neither will technology emerge as a savior in relation to any of the severe sustainability challenges of today, nor will it be possible to ignore technological innovation or will it make sense to neglect technology as an instrument to improve socio-technical systems (e.g. renewable energy technologies, more efficient fuels and drive technologies in the transport sector etc.). The evolutionary perspective on which the MLP is based has also been criticized at a more fundamental ontological level. Concepts of variation and selection are criticized for unduly simplifying the basic mechanisms involved in social processes of change. Selection is treated as a process that functions independent of deliberate human action and thus the conditions for change are conceptualized as a factor that is external to the actual process of transformation. Actors are relatively passive, reacting to system shocks or crises. The resulting transformation processes are further conceptualized as shifts between phases of equilibrium and it is assumed that the different phases of stability and radical change can be clearly separated. Proponents of relational approaches based on flat ontologies (Geels, 2010, p.507), e.g. Actor Network Theory (cf. Latour, 1987) or Social Construction of Technology 93

107 approaches (Bijker et al., 1987), argue that the focus on external factors and selection environments is overly structural and denies the role of agency in transformation processes. It is assumed that actors and technological artifacts and infrastructures are interrelated in flat networks rather than structural hierarchies and that actors are engaged in processes of negotiation and deliberation. Transitions are then conceptualized and studied as outcomes of such deliberative processes rather than macro-level phenomena of shifting equilibria (Cooke, 2011, p. 109; Garud & Gehman, 2012, p. 983 f.; Geels, 2010, p. 502; Geerlings et al., 2009, p. 410). Such a fundamental critique of evolutionary approaches is important, in order to raise awareness for the basic assumptions inherent in evolutionary thinking, especially when applied in the study of socio-technical sustainability transitions: From an evolutionary perspective, the image of a journey is one of making progress through technological disruptions of moving forward while battling the remnants of the past (Garud & Gehmann, 2012, p. 986). In view of current sustainability challenges, traditional concepts of progress, achieved through technological innovation and economic growth, may be counterproductive. It is therefore important in the analysis of sustainability transitions to recognize and be explicit about the normative diversity involved and its theoretical as well as practical implications. For theoretical analyses this means that frameworks should rather be open with regard to complexity and diversity in transition processes, embracing evaluative diversity and recognizing divergent normative vectors (Stirling, 2011, p. 84). This means a partial deviation from innovation theory, as regards some of its basic premises of onedimensional economic and technological progress, and instead awarding a more prominent role to agency and societal deliberation processes as well as social innovations or social practices that do not fit a primarily technological logic of innovation as progress (Garud & Gehmann, 2012, p. 986; Koch & Grünhagen, 2010, p. 234; Stirling, 2011, p. 83 f.; Volkmann & Tokarski, 2010, p. 154). A weak perspective on agency and the politics of transitions is a general criticism of the MLP and it has been shown that this can, at least partly, be explained by the basic assumptions of evolutionary thinking inherent in innovation studies. However, when reflecting on the sociological concepts that also feed into the MLP and the normative diversities inherent in almost any theoretical approach, it can be argued that this conceptual weakness is just as much a problem of operationalization and application. Despite the seemingly structuralist MLP graph (see Fig. 8), the role of human agency in transitions is explicitly addressed by referring to structuration theory and the role of knowledgeable actors 94

108 as well as a focus on niche-regime dynamics with a view to power constellations among concrete actors and by clearly distinguishing levels of structuration in social practices, rather than grouping actors in fixed niche or regime categories. In transition studies in general, a tendency can be observed to foreground the role of technology as the central element around which transitions evolve more or less mechanically and where the complex dynamics involved are unduly simplified, e.g. by applying the MLP in a rather superficial way (de Haan & Rotmans, 2011, p. 91). Especially for the purpose of this thesis, studying e-mobility as a potential system innovation rather than diffusion processes of electric vehicles such an oversimplification needs to be avoided. It is therefore important to be clear about what the MLP is, and is not, in terms of its specific explanatory style and research approach. The MLP can be grouped among middle-range theories that offer heuristic frameworks rather than precise models, which aim at explaining causal links between predefined variables. The MLP instead focuses on processes, complex dynamics and contingency, in order to explain transitions (Geels, 2011, p. 34 f.). It is thus well suited for approaching research questions such as the one posed in this thesis. The potential for e- mobility to develop as a sustainable system innovation depends on various influencing factors that will become relevant only in mutual types of interactions and in the context of emergent process sequences, which makes it hardly possible to identify a manageable selection of independent variables and clearly hypothesize what their relation is to a narrowly defined dependent variable. The concept of system innovation or transition captures complex and structural change across different societal domains, which makes it difficult to carry out a variable-oriented analysis in any meaningful way. Due to this empirical complexity, an open framework like the MLP may be better suited, with the consequence though that the MLP is not a truth machine that automatically produces the right answers when the analyst enters the data (Geels, 2012, p. 474). Furthermore, this means that, as regards methods, narrative explanations play an important role. Since these always include elements of interpretation by the analyst, they have to be carried out in a transparent and detailed way, in order to ensure scientific rigor. The difficulties involved in applying such methods should not lead to their outright rejection, since they can contribute to a more profound understanding of complex phenomena such as transitions (Geels & Schot, 2010, p. 101). However, these types of overarching frameworks are generally difficult to operationalize and apply to empirical cases in meaningful ways. No straightforward hypotheses can be built and delineating the object of study as well as tracing influencing factors and multi-level dynamics is rather part of the research question, than premise on 95

109 which the research is built. The strength of the MLP lies thus in identifying the relevant questions and opening up new perspectives, rather than prescribing a model explaining concrete transition dynamics for specific empirical cases. The MLP has been developed as a sort of boundary object where many disciplines can relate to each other and communicate their different understandings and concepts of transformation processes (Grin et al., 2010, p. 4). It functions as a broad heuristic aiming at a comprehensive understanding and exploring of transitions as a whole. This explains why the MLP is crafted as a macro-level framework but it also shows that using it as a mechanistic and structuralistic theory is a misuse. It serves as a guiding basic framework, which allows, and requires, the analyst to zoom in on actors (Geels & Schot, 2007, p. 414) at the micro-level and in concrete empirical cases. Even though the MLP graph depicts structural levels and basic dynamics, these are the outcome of social practices, deliberation processes and power struggles among actor groups (ibid.). Thus, the usefulness of an MLP analysis will in the end have to be determined empirically. Despite a continuously growing body of case studies and similar applications of the MLP, issues of operationalization and empirical validation of the regime concept are widely debated. General insights are that the MLP s central concept of regime does not per se prescribe a specific scope or unit of analysis. Drawing the boundaries of a specific regime or socio-technical system will have to be determined in relation to the societal function that is in the focus of the respective study and in relation to the specific research question. However, this should be done in a clear and explicit fashion and by pointing out the analytical strength of the regime concept for a specific case. Such methodological openness and clarity is lacking in many empirical case studies so far (Geels, 2011, p. 31; Holtz et al., 2008, p. 624 ff.; Markard & Truffer, 2008, p. 606 f.). Apart from defining the correct way of applying different conceptual levels empirically, it seems important to keep in mind that transition studies are essentially part of a critical social science tradition and aim at [engaging] subjects in processes of reflection in ways that feed back into their practices, contributing to a re-orientation towards sustainable development (Smith et al., 2010, p. 445). Thus, with regard to empirical operationalization it may also make sense to focus on regimes not only as a specific conceptual level in the MLP, but also as a relevant notion regarding the structuration of local practices (Shove & Walker, 2010, p. 474; Geels, 2011, p. 37). Reflexivity with regard to theoretical concepts and sensitivity with regard to the respective empirical case is needed, in order to question the meaning and significance to action of levels on the ground, from the perspective of subjects (Genus & Coles, 2008, p. 17). Transition studies are thus faced with tensions between aiming 96

110 at becoming established as a field of research by the standards of the science system (publications, citations, quantitative methods, scientific rigor based on traditional standards) and its inherent desire to have practical implications, to act as transformative science (cf. Schneidewind et al., 2015 forthcoming; Schneidewind & Singer-Brodowski, 2013; Schneidewind & Augenstein, 2012). A related difficulty is that the MLP is mainly applied ex-post, explaining historical cases of transitions. In these cases, starting and end points of transitions can be identified and the dynamics and patterns emerging in the process can be traced. Ongoing transitions, or desired transitions towards sustainability are thus much more difficult to be studied. The following chapter therefore deals with the question of whether and how it is possible to apply the MLP not only ex-post but also in a (potential) ex-ante situation. Two related frameworks, transition management and strategic niche management, will be discussed as examples of more future-oriented approaches with a focus on shaping and steering transition processes. And it will be asked whether the MLP-framework can be adapted in such a way that it fosters a better understanding of current dynamics whether or not these amount to an actual, full-blown transition eventually. 3.2 Analyzing Transitions in the Making Transition studies are motivated by a concern for sustainable development and are thus inherently future-oriented. The underlying research question from which this field of research emerged is, basically, how to facilitate transitions towards future sustainability? It has been recognized that sustainability challenges in fields such as energy, transport, water or food can only be met by achieving structural transformation processes in complex societal systems. These challenges are approached scientifically by referring to established theories and fields of study, e.g. evolutionary economics, science and technology studies, complex systems theory or sociological grand theories, in order to learn about the phenomenon of transformation in general. The MLP has emerged as an overarching heuristic for analyzing past transitions, identifying general patterns and mechanisms of change in socio-technical systems. Based on the MLP, orientation towards the future comes in the form of a focus on transitions in the making and is spelled out in two prominent approaches in the field of transition studies, namely Strategic Niche Management (SNM) and Transition Management (TM). 97

111 Strategic Niche Management is an approach that focuses on fostering niches as the central nucleus for radical novelty from which transitions evolve, while Transition Management entails a broader perspective on the governance of transition processes across the different levels and phases of the MLP. Governance for sustainability transitions is conceptualized as a complex and long-term process. Assuming this type of complexity, and despite the fact that the term management in SNM and TM suggests otherwise, it is argued that transitions cannot be managed in any straightforward manner, or stimulated by traditional command-and-control policies at least not in a simple way and so all one can do and hope for is to exercise some influence, or leverage, to modulate ongoing dynamics (Hoogma et al., 2002, p. 198). Strategic Niche Management and Transition Management are therefore not straightforward management tools, but rather approaches that are based on the premise of such complexity and which can serve to guide interventions in a dynamic context. In that sense, transition management and strategic niche management are portrayed as governance in themselves (Nilsson et al., 2012, p. 52 f.). So, even though research on technological innovations, innovation systems, and science and technology studies have been a major influence on transition studies, policy and political processes have become central issues, especially when the focus is on envisaged future transitions to more sustainable energy or transport systems. The broad definition of socio-technical systems as the basic unit of analysis of transition studies necessarily includes policy regimes as a central type of sub-regime characterizing the overall configuration of any specific socio-technical system. The central delineation criterion of such a system, i.e. the fulfilling of a generic societal function, also points to the central role of the political sphere, because this is where societal needs, problem agendas and preferences as well as instruments for achieving societal goals are negotiated and implemented (Kern & Howlett, 2009, p. 392). Thus, public policy has been identified as a key factor for influencing transitions and therefore SNM and TM have been developed as policy approaches suitable for dealing with complex issues of sustainable development. The basic challenge that policy in this field is faced with is uncertainty with regard to the complexity of sustainability challenges, the long-term and dynamic nature of change processes, and the resulting array of questions emerging from these framework conditions when aspiring to design policies and policy instruments for shaping the future. The basic premise from which SNM and TM have been developed is to deal with this type of uncertainty. Therefore, both concepts include a focus on a desired future, while keeping concrete instruments adaptable and open to reflexive processes of re-adjustment. 98

112 Another challenge that SNM and TM aim to address is to deal with conflicts emerging between bottom-up initiatives and radical novelty and established actors representing vested interests and using their power to hamper efforts towards sustainability transitions (Voß et al., 2009, p. 281). Referring back to the MLP, the central rationale of TM and SNM as futureoriented approaches for fostering sustainability transitions is to facilitate niche-regime dynamics. It is assumed that the seed for change emerges in niches and that transformative dynamics need to move to the regime level, because, from a pessimistic point of view, they aim at changing currently unsustainable regimes that are categorically and fiercely opposed to any kind of change, and, from an optimistic point of view, only powerful and resourceful regime actors have the capacity to up-scale niche innovations and to significantly contribute to systemic change. As shown by the different transition patterns, in reality there will be variations with regard to the relations between actors within and across niches and (sub- )regimes. SNM and TM deal with these processes, in practice and in theory, by identifying niches with a potential for influencing regime structures, studying ways of improving niche development and fostering processes of up-scaling. With regard to the overall question of this chapter how to deal with transitions in the making? SNM and TM both depart from the premise that the potential for system innovations depends on well-developed niches and the way that they successfully manage to influence regime structures 4. The basic aim is to analyze and understand such processes and, based on this, to develop suggestions and instruments for improving up-scaling in practice Strategic Niche Management (SNM) Strategic Niche Management is an approach that aims at developing and nurturing protected environments, i.e. niches, in which new technologies and alternative social practices can flourish and grow, eventually serving as the starting point for regime change. This type of protection is needed, because radical technological or social innovations can hardly emerge when they are from the beginning subject to adversary selection pressures by the established regime, i.e. if they are in sharp contrast to dominant rules and guiding principles. Niches are therefore needed as a protected space in which radical alternatives can mature and eventually expand to replace the existing regime (Markard et al, 2012, p. 957 f.; Wells & Nieuwenhuis, 2012, p. 1683). In view of this intended expansion, it is also important not to over-protect, but A similar discussion of the concepts of SNM and TM has been published in a co-authored conference paper: Sterk, Augenstein & Wehnert (2014). 99

113 to open up the niche step-wise. This can be termed selective exposure and illustrates the interrelations between niche and regime: Strategic niche management is a process in which a new technology and the existing environment can adjust to each other, because the creation of the niche changes the selection environment (Schot et al., 1994, p. 1073). Three distinct processes have been identified in the literature that characterize wellfunctioning niches and thus are important elements addressed by SNM. First, social learning processes are taking place in a niche, e.g. regarding the application of a specific technology in real-life circumstances, changes in user behavior, the necessary framework conditions, useful policy instruments, or the shaping of symbolic meaning of specific technologies or practices. Second, expectations and visions are developed in a niche. This serves two purposes, namely providing guidance for niche-internal processes, and attracting niche-external interest and support. Third, heterogeneous networks are built within and around a niche. These can be important for the envisaged expansion of a niche into a regime, if the network includes also regime actors and provides access to resources (Kemp et al., 1998, p. 190 f.; Hoogma et al. 2002, p. 28 f.; Geels, 2012, p. 472; Markard et al., 2012, p. 957 f.). Overall, SNM can be characterized as a bottom-up approach and in practice often takes the form of pilot or demonstration projects. Proponents of SNM argue that niches should from the beginning rather be conceptualized as experiments instead, because SNM is less about testing technologies and user acceptance, but rather focuses on broader learning processes, where pre-existing user behavior, the way technologies are used and how they are embedded in their social context may be challenged and maybe even change (Hoogma et al., 2002, p. 5; Markard et al., 2012, p. 957 f.). Thus, SNM as an approach aims at developing niches in such a way that learning at this deeper, more structural level becomes feasible. The basic rationale for such an approach is based on the assumption that transitions are processes of socio-technical co-evolution, which are not a pre-determined outcome of the diffusion of any specific technology and socially more complex than market mechanisms. Therefore, experiments in niches should be designed as open-ended search and learning processes (Hoogma et al., 2002, p. 4). In this respect, SNM is understood to be a new governance approach that goes beyond focusing on specific policy instruments aimed at achieving clear-cut goals or fostering specific green technologies. It stimulates interaction between different actor groups and is intended as a way to proactively deal with conflicts and dilemmas emerging in striving for sustainable development. Thus it can basically be understood as an approach of reflexive governance where the involved actors learn to deal with complex change processes in a step- 100

114 wise, deliberative and experimental way. This will allow for fundamentally new sociotechnical system configurations to be tried out in niches and, by including as many of the relevant actors as possible, for finding ways of up-scaling these niche innovations to the regime level (Hoogma et al., 2002, p. 181 f.; van den Bergh et al., 2011, p. 13). Practical experience based on projects that have from the beginning been designed on principles of SNM so far remain limited, because the concept is still relatively young. However, Hoogma et al. (2002) have analyzed e-mobility projects carried out throughout Europe during the 1990s based on criteria of SNM. These cases have shown that a number of aspects are of crucial importance for successful SNM. First, it is important to clarify from the start the experimental nature of projects. They should be designed in such a way that underlying assumptions, e.g. regarding suitable technological options, the right strategies for diffusion and implementation, and normal user behavior, can be questioned and challenged. In practice this can be achieved by defining the concrete goals of projects in suitable ways. They should be formulated in terms of qualitative learning goals, rather than quantitative goals of diffusion rates or market shares. Furthermore, specifying goals should be done in a reflexive way, i.e. re-adjusting concrete goals throughout the process. Second, since SNM entails deliberation and navigating conflicts across a wide range of actor groups, setting up the niche with regard to the involved partners needs to be done carefully. In order to secure the necessary commitment and willingness to cooperate of all partners, it is suggested that expectations should be articulated openly and repeatedly throughout the course of a project. Finally, with regard to the up-scaling of niche innovations, a cautious and foresighted attitude is imperative. Barriers and opposition at the regime level should be anticipated and assessed realistically, in order to deal with them proactively and choose a suitable, often gradual, up-scaling strategy (Hoogma et al., 2002, p. 197). Apart from the practical implementation of SNM, the conceptual approach itself has also been criticized. One major point of critique, which applies to other approaches in transition studies as well, is the often limited focus on future change (Wells & Nieuwenhuis, 2012, p. 1683). Such a focus remains limited when it fails to simultaneously take sufficient account of the barriers precluding this envisaged future change, while instead focusing, for instance, exclusively on managing processes within niches. Furthermore, the concept of SNM sometimes tends to be power-blind. It explicitly advocates including a broad range of actors, where useful also industry actors with vested interests in preserving the dominant regime. This may lead to challenges in practice, when powerful incumbents dominate 101

115 concrete projects. Especially in the context of sustainability transitions, the aim is radical change and this requires a careful balance of resourceful incumbents and niche actors with new ideas and alternative approaches (Hoogma et al., 2002, p. 201). A major challenge, in theory and in practice, is thus to place SNM in the wider context of governance for sustainable development. SNM is based on a policy approach that, against the background of uncertainty involved in achieving sustainability goals, focuses on modulating on-going dynamics and utilizing windows of opportunity (Hoogma et al., 2002, p. 200). However, this type of policy approach cannot replace more traditional policy approaches and instruments, e.g. those that control pollution, set concrete boundaries and incentives, and thus translate and transfer sustainability objectives into established policymaking structures, which often follow an economic or market logic. SNM is neither a useful substitute for these kinds of policies, nor a way to improve them, but rather an addition that helps to deal with underlying questions and produces a broader range of alternative (technological) options (Hoogma et al., 2002, p. 200) Transition Management (TM) Transition Management is another central approach that has emerged from the search for new forms of policy design that are more reflexive, embedded in complex system realities and geared towards a long-term future (Voß et al., 2009, p. 276). Similar to SNM, the aim is not to replace existing policies, but to add new elements that are better able to deal with the uncertainties inherent in the process of fostering sustainable development in key societal domains. In contrast to SNM, the approach of Transition Management is more comprehensive and has a broader focus on transition processes as a whole. TM has been developed together with politicians and scientists and was adopted by the Dutch Ministry of Economic Affairs as an overall policy framework in It proceeds along five cyclical steps: First, a transition arena is established, involving all relevant stakeholders for a specific policy problem. The aim of bringing those actors together is to induce processes of second-order learning by facilitating a common understanding of the underlying problem, the different perspectives held by others and the resulting interests and strategies. Second, this should form the basis for developing a shared vision among the involved actors. In a deliberative process, the basic goal of achieving, for instance, a more sustainable energy or transport system, is spelled out more clearly. The resulting vision should ideally be shared by all involved actors and serves as an orientation for the following steps. Third, and based on the particular vision that has been developed, possible pathways 102

116 that might lead towards this envisaged future are being developed. Fourth, concrete experiments are designed and carried out in practice, in order to test options that have been identified to be promising. These experiments can and should be ambitious, in terms of going beyond established routines and policy instruments. They can be first steps on the pathways developed at earlier stages of the TM process. As a final step, the experiments should be monitored and evaluated, in order to facilitate a thorough learning process. This should result in continuous re-adjustment and re-evaluation of experiments, in order to find ways of upscaling these experiments to the regime level. The different phases of TM are not clearly separated in practice, there are often feedback-loops and interlinkages across the different steps. For instance, experiments can induce a process of going back to earlier stages of TM, because insights that have been gained in practice may facilitate re-adjustments of future visions or potential pathways. It may also show that other actors should be included that were not considered relevant in the beginning. This shows that, similar to SNM, the basic aim is to establish processes and settings that facilitate learning processes and reflexivity in the design of concrete experiments, rather than implementing clear-cut instruments for achieving predetermined quantitative goals (Loorbach, 2010, p. 171 ff.; van den Bergh et al., 2011, p. 13; Voß et al., 2009, p. 284 f.). TM can thus be characterized as a new form of reflexive long-term policy. The way TM is conceptualized expresses skepticism with regard to older planning approaches and naïve ideas of controlling and managing (in the narrow sense of the word) processes of change. Instead, the TM approach combines an orientation toward a long-term vision of sustainable development with short-term experimental learning to probe options and find pathways to realize the vision (Voß et al., p. 277). This is done in a cyclical and recursive process, where the vision and the experiments are continuously reflected upon and if necessary re-adjusted. This requires a certain open-mindedness with regard to the definition of concrete aims and policy instruments, in order to be able to change them throughout the process (Voß et al., 2009, p. 277). Similar to SNM, the basic rationale behind this approach is the idea to modulate, rather than interfere in, dynamics that are already shaping a sociotechnical system, because [t]he complexity of the system is at odds with the formulation of specific, quantitative objectives (Voß et al., 2009, p. 284). Accordingly, TM rather aims at facilitating the context in which this becomes feasible, instead of focusing on fixed goals and rigid strategies. In the TM approach and the way it has evolved, struggles of dealing with the nicheregime interface and up-scaling of experiments are apparent as well. The rationale for setting 103

117 up transition arenas is to provide a space for developing alternative ideas and practices and shielding them from regime pressure. Therefore, regime actors were intentionally excluded in early versions of the TM concept. In subsequent steps of refining the TM approach, it was argued that regime actors should indeed be included in transition arenas, because their financial and non-financial resources, e.g. societal legitimacy and political standing, could aid the process. This makes sense, if the eventual goal is to achieve a broader up-scaling of niche experiments. At the same time, this poses a threat of niches being captured by regime actors and experiments being steered in ways that impede the original goals of radical change. Ideally, there should be pressure on existing regimes, e.g. in the form of environmental laws or taxes, and against this background diverse niche developments would be granted the chance to flourish and compete with each other, in the end resulting in an evolutionary process of successful niches transforming regime structures (Kern & Smith, 2008, p. 4094). However, despite the fact that transition management has been developed in a practical policy context in the Netherlands, it has so far remained predominantly an academic approach. It can be shown that on a conceptual level, issues of power in political processes and institutions tend to be ignored, or at least are not considered sufficiently. This theoretical blind spot leads to weaknesses and unintended outcomes in the implementation of TM processes. A number of studies show that in some cases incumbent actors have managed to capture TM projects and sustainability discourses, in order to hamper the achievement of substantial sustainability-oriented goals (Avelino, 2009, p. 381; Meadowcroft, 2009, p. 336; Voß et al., 2009, p. 277 ff.). Thus, problems with the implementation of TM in practice often have to do with a failure to account for political realities. Since TM is designed as a very open framework for deliberating and exploring options with large stakeholder constituencies, there is a threat that the process is captured by powerful regime incumbents defending their vested interests. The fact that these actors are usually more experienced and resourceful as compared to other interest groups, e.g. civil society organizations, is often not accounted for by the process design of TM. Similarly, by focusing on TM as a way of facilitating open-ended learning processes, the focus on concrete sustainability challenges and an awareness of the political struggles involved are sometimes lost (Avelino 2009, p. 381 ff.; Voß et al., 2009, p. 286 f.). Another major challenge is the establishment of TM as a new policy approach itself. As is the case with SNM, TM is understood to be not a substitute or element of existing policy instruments, but rather as an additional or cross-cutting policy approach. However, this means that TM needs to be established within the context of existing and long-established 104

118 political processes and routines. TM is a reflexive approach oriented towards the long term and this can be in conflict with existing patterns of policy-making, which may in turn lead to unintended effects when it comes to implementing TM. For instance, in the implementation process it can be observed how substantial sustainability-oriented goals are being reduced to more conventional measures of technological development and fostering economic competitiveness (Avelino 2009, p. 388; Heiskanen et al. 2009, p. 424 f.; Meadowcroft 2009, p. 329 f.). Since these [c]ontextual dynamics feed back into the design process (Voß et al., 2009, p. 289), TM itself can be seen as an experiment that needs to be constantly and critically evaluated and re-adjusted Fostering Transitions in the Making through SNM and TM? Both SNM and TM have been developed out of a concern for designing policies that are better able to foster sustainable development. Dealing with sustainability challenges requires a perspective on the long term, an integrated perspective on interlinked societal domains, and on system innovations, i.e. the emerging interactions between technological innovations and social practices. Thus, the policy agenda for sustainable development basically aims at structural change in and of societies. Established political processes, such as electoral cycles spanning four or five years, political institutions, such as the traditions and culture passed on across generations of civil servants, or conventional policy instruments, which are focused on specific sectors and shorter time spans, are hampering the development of policies that can meaningfully address sustainability problems. SNM and TM are approaches that are specifically designed with the aim of fostering structural change, for instance by focusing on socio-technical systems fulfilling generic societal functions, rather than on specific sectors or technologies. This integrated perspective is supplemented by a long-term orientation, assuming that transitions generally take 25 to 30 years. Recognizing that sustainability problems are complex and intertwined with change processes unfolding over substantial periods of time, leads to a substantially different approach when it comes to designing policy. It can be compared to the shifting focus within political sciences from government to governance, because SNM and TM are processoriented with a focus on a diverse set of relevant actors and their interlinkages. Potential for triggering structural change is not seen in specific policy instruments, but rather in the way that ongoing dynamics are utilized by a reflexive governance approach, where specific instruments are constantly evaluated and adjusted in flexible ways. SNM, and especially TM, rely on a mix of bottom-up and top-down approaches, i.e. weakening established structures 105

119 where possible and creating spaces where alternatives can develop. In essence, SNM and TM are providing frameworks for reflexive social learning processes, in order to deal with the difficulty of aligning long-term visions of sustainability with the need for concrete action in the short term (van den Bergh et al., 2011, p. 13; Voß et al., 2009, p. 278, 281). However, despite these noble aspirations, SNM and TM are faced with some major challenges conceptually and in practice. Most importantly, it is ignored that the political sphere is itself part of the dominant regimes in unsustainable socio-technical systems and that public policy is not an exogenous element that can introduce drivers for change from a neutral position. It is thus not sufficient to develop new concepts for more sustainability-oriented policy-making. A thorough analysis of the political domain itself and of the way that problems inherent in processes of policy-making can be addressed by concepts such as SNM and TM, is needed (Kern, 2012, p. 300; van den Bergh et al., 2011, p. 13 f.). A related challenge has to do with the process-orientation of SNM and TM. While this is important, considering the complexity and duration of the envisaged change processes, it also entails the danger of losing direction. Achieving sustainability goals is the basic motivation, which should not get lost by focusing solely on process design. Concrete targets and suitable instruments to reach them should in some way be defined and guide the process (Voß et al., 2009, p. 298 f.). This is important especially because the very open processes in SNM and TM in combination with the principle of including regime actors in transition arenas, in order to eventually cross the boundary between niche and regime, may lead to unintended outcomes. Powerful regime actors may capture niches and actively prevent system innovation by using their influence and resources to shape them as market niches where technologies or business models are developed that follow established regime logics or guiding principles such as immediate economic feasibility and market success (Voß et al., 2009, p. 286 ff.). Experiences with the Dutch energy transition, where TM has been developed, have shown that even in a situation where transition arenas had been set up, ambitious goals had been agreed on and the process was initiated in a generally open atmosphere with regard to innovative policy approaches, the expected outcomes did not materialize. Established political institutions and processes could not simply be overridden or bypassed and presented formidable barriers precluding the emergence of system innovations. This implies that in the field of transition studies, and SNM and TM in particular, analysis for policy needs to be complemented by analysis of transition policies and their politics (Kern & Smith, 2008, p. 4101). 106

120 3.2.4 Transitions in the Making : How to Analyze the Potential for System Innovation ex ante? Approaches such as SNM and TM are important for inducing open-ended social learning processes, relying on reflexivity and adaptability as operational principles in view of a longterm orientation, in order to facilitate sustainability transitions. However, the critique of SNM and TM has also shown that they need to be accompanied by an analysis or monitoring of the actual potential for transformative change to emerge at different stages of transitions in the making. This cannot be achieved by focusing on principles of SNM and TM as sole analytical concepts, because they focus on procedures and design of processes for new forms of policy. Apart from that, not all developments or policies that may have a potential for system innovation are from the outset designed as SNM or TM processes, as e.g. in the context of the Dutch energy transition. An assessment of potential in an ex ante situation, i.e. in the case of a presumed transition currently in the making, requires a broader perspective on systemic change that goes beyond a focus on process and looks at niche-regime dynamics with a view to the involved power struggles and political processes. This is important because an analysis of a specific niche carried out along the principles of SNM, such as the emergence of shared visions and broad network-building, may result in findings of potential for system innovation that turns out to be at least questionable. For instance, Geels (2012, p. 472 f.) argues that the development of alternative vehicles, battery-electric as well as fuel cell vehicles, presents a case of successful SNM. These technologies have been developed in niches, visions, in terms of expectations related to these technologies, have become broadly shared, and eventually broad networks have emerged where car manufacturers have bought or formed alliances with smaller companies active in the field of alternative drive technologies. According to the process-oriented principles of SNM this may indeed be a successful case. However, from a broader transition perspective, this may only be true at first glance. It is at least doubtful whether this is a case of structural change with a potential for system innovation that radically alters established regime structures. Much rather, this seems to be a case of a niche being captured by regime actors. Even if alternative drive technologies have been developed in niches guided by a broader vision of sustainable mobility and radical change, car manufacturers are stepping in and taking over these new technologies to integrate them in their normal activities of producing and selling cars without major changes to their basic strategies or business models. They have vested interests in preserving the current transport system and by entering the niche of 107

121 alternative drive technologies and vehicles, they have found a way to steer developments in directions that do not threaten these interests. At the same time, the development of alternative drive technologies at a larger scale only becomes possible with the help of these large car manufacturers expertise and financial resources in the field of R&D. This is the very basic dilemma with regard to the niche-regime interface, that on the one hand, resourceful regime incumbents are needed for a broad upscaling of experiments, while on the other hand, their involvement usually precludes the emergence of radical system innovations. In contrast to this being presented as a questionable example of successful SNM, there are indeed some valuable lessons to be learned from SNM and TM with regard to the potential for system innovations emerging from transitions in the making. As argued by Hoogma et al. (2002, p. 181 f.), SNM does not aim primarily at the development of specific technologies or network-building per se. It is in essence about learning processes at the level of rules and guiding principles, questioning normal behavior and routines. For the case of electric vehicle development, successful SNM would induce this type of learning, rather than technology diffusion (which is part of the process but not the central outcome ). In the beginning, an electric vehicle niche would most likely be characterized by established rules and guiding principles. In practice this could mean that electric vehicles are expected to function much like conventional cars and they would be assessed and tested against the familiar concept of personal automobility with ICE cars. The learning processes envisioned in SNM would then lead to a stepwise departure from these structures, for instance, when users begin questioning their mobility behavior and possibly develop new routines, or car manufacturers considering new business models in cooperation with energy providers, who may start thinking about new forms of energy services. Thus, the learning processes relevant for successful SNM are broad, covering many issues and actors, and deep, going beyond established routines and guiding principles (Schot et al., 1994, p f.). Fundamentally, SNM is about the co-evolution of a technology and its environment at a deep-structural level. This has consequences for the evaluation of whether or not an SNM project has been successful. Success does not so much depend on direct economic success or the emergence of marketable products as is well illustrated by the example of new drive technologies taken up by large car manufacturers, which is most likely not contributing directly to the emergence of a system innovation. The outcome of social learning processes are a more relevant success in this regard, but also much harder to measure. It has been proposed to focus on expectations and visions of the future that are developed in the context 108

122 of an SNM project. Changes in images of a desired future and new guiding principles emerging from this can influence decision-makers in politics and industry based on what has been learned in niche experiments. This may eventually have a real and visible effect, when it translates into concrete strategies and behavior (Hoogma et al., 2002, p. 195 f.). Following from this, it is important to note that therefore also failures and problems emerging in the context of such experiments can amount to valuable lessons. During early phases of transitions or experimenting with alternative practices there will inevitably be failed attempts and approaches that turn out to be unsuccessful. This may not necessarily cast doubt on the experiment in general, but could just as well be an indicator for the premature state of a niche (Hoogma et al., 2002, p. 196). In order to learn about the potential for a system innovation to emerge, it is thus all the more important to also analyze cases of unsuccessful niches or experiments, which is often neglected in transition studies. Especially because efforts of fostering transitions take place in established policy contexts and against the background of dominant regime structures and this entails a danger of poor implementation and outcomes (as discussed earlier), it is important to identify and analyze cases of failed efforts and the barriers impeding system innovation (Kern & Howlett, 2009, p. 392 f.). So, how can cases with a potential for emerging system innovations be analyzed? Principles of SNM and TM alone will not suffice, because they primarily focus on process design. What can be learned from these concepts is that the analysis should not focus on superficial criteria, e.g. diffusion rates or economic success of a new technology, but rather on deep-structural learning effects relating to dominant rules and resources. An important aspect is also the niche-regime- interface, especially a cautious view to power relations, the politics involved and possible dynamics of niche capture. These aspects are central in the MLP (which has of course also been the basis for developing concepts of SNM and TM) and recently, approaches have emerged that use the MLP as an analytical tool for assessing, for instance, policies that are aimed at fostering transitions. A prominent example of such an approach has recently been presented in a paper by Kern (2012) where the MLP is explored as a framework for evaluating a UK policy initiative ( Carbon Trust ) that aims at a transition to a low carbon economy. This policy initiative has been selected as a case study applying the MLP as a tool in the analysis of policy, rather than for policy. Applying the MLP as an analytical framework for ex ante assessment of policy builds on the insight that governments and public policy play a major role in transition studies, while it is often unclear what this means in terms of concrete instruments and processes, especially when considering that transition policies are not developing in a void 109

123 but in the context of established political structures. Therefore, the MLP is used here as a framework for analyzing in how far existing policies may or may not stimulate transitions. This approach departs from traditional applications of the MLP to historic cases of transitions and turns towards envisioned transitions of the future by ex ante analyzing the policies that might contribute to realizing these future visions (Kern, 2012, p. 299). Similar to the focus on learning in SNM, here the focus ought to be on the deep-structural impact of policy, rather than a quantitative assessment of short-term, direct and easily measurable outcomes of a specific policy. Consequently, the basic concepts of the MLP are operationalized in a rather general fashion, focusing on learning processes at the niche level, changes in rules at the regime level, and on macro-economic trends representing the landscape. It is then analyzed in how far the selected example of a low carbon policy initiative contributes to the development of a wellfunctioning niche, in terms of improving specific low carbon technologies but also broader capacity building in networks, and changes at the regime level, again looking at diffusion of technologies but also social learning processes in established networks. The landscape level cannot be influenced by a single national policy initiative, the analysis here focuses on the exogenous drivers and barriers for policies aiming at transitions. Landscape trends are thus analyzed in terms of their impact on the specific policy being studied and it is shown how policy interacts with broader macro-economic trends and, for instance, is related to cultural values or philosophies of policy-making (Kern, 2012, p. 300 ff.). Nonetheless, it will not be possible to predict the outcome of a specific policy, simply due to the complexity involved in envisioned low carbon transitions, even when the ambition is narrowed down to only focus on energy or transport systems which are still highly complex in themselves, including many actor groups, political domains, technologies and infrastructures. Adding a long-term perspective, complexity increases even more. This realworld complexity is reflected in the MLP, which refrains from explaining transitions as phenomena of direct, linear causalities with clearly identifiable variables driving the process independently. Rather, it is acknowledged that transitions are always an outcome of contingent dynamics and interlinkages of a multitude of variables and processes (Geels, 2012, p. 474; Meadowcroft, 2009, p. 328 f.). To get a grip on this complexity, historical cases have been studied, in order to identify characteristic patterns and mechanisms playing out in transitions (Geels & Schot 2007, as discussed in ch ). While it has been criticized that this type of reconstruction is not scientifically sound enough in terms of validity (see Dolata 2013, p. 27), it has produced insights in past events that do pass a plausibility-check 110

124 assuming that due to the inherent contingency and highly complex nature of transition processes a greater degree of validity is possibly not feasible at all, this has been an effective way of scientifically reducing complexity. Even if this approach is useful for understanding past transitions, and maybe produces knowledge for the present with regard to specific patterns or mechanisms observable over limited time frames, scale and scope, it still does in no way allow for predicting the future: Looking back from a vantage point in 2100, analysts may be able to discern the start and end of each distinct transition and identify the abortive efforts and reversals that characterize the overall movement. But for those living through these processes, it will be much more difficult to understand what is really going on (Meadowcroft, 2009, p. 329). Nonetheless, as shown by Kern (2012), generic patterns of niche-regime dynamics or the impact of landscape factors can help inform ex ante analyses of policies intended to stimulate transitions. The aim of such an analysis is certainly not to predict the future, as for instance compared to scenario methods, but it can provide insights into whether or not efforts are actually facilitating dynamics in the desired direction, which is an aspect sometimes not reflected enough in SNM or TM. Thus, the paper by Kern makes a valuable contribution by showing how the MLP is used in a novel way: as a heuristic to ex ante assess policies ( ) [and] their likely impact against the background of theorizing about patterns of large scale, socio-technical change (Kern, 2012, p. 298). Assuming that it is not possible to precisely predict policy outcomes, and acknowledging that even ex-post policy analysis can sometimes hardly identify outcomes and unavoidably remains limited to outputs (again due to real-world complexity and non-linear causalities), an MLP-based ex ante analysis can still be useful. Policies can be evaluated against the background of ideal-typical multi-level dynamics leading to transitions in theory, e.g. regimes need to be destabilized, niches nurtured and landscape factors need to be taken into account. This will lead to an assessment of the degree to which a policy initiative contributes to such dynamics, thus an analysis of a policy initiative s contribution to the potential for a transition to emerge: how well is the particular policy initiative suited to stimulate structural change, while always keeping in mind that the eventual outcome with regard to an overall transition of a socio-technical system as a whole depends on a plethora of other factors as well. Even in the face of such complexity and uncertainty, this type of ex ante assessment has advantages as compared to more classic evaluation methods based on quantitative cost effectiveness assessments or measuring reduced emissions (Kern, 2012, p. 308). 111

125 This is particularly relevant for a case such as e-mobility. A quantitative assessment of diffusion rates or reduced emissions would not say anything about the system-innovative potential inherent in the development of e-mobility. What is more, it would most likely not produce positive results and could even trigger efforts going in a completely wrong direction, from a sustainability point of view at least. Looking at the slow diffusion of electric vehicles and their inability to compete with conventional cars, which would be important indicators in a quantitative assessment of policies fostering e-mobility, the consequence could be either discontinuing e-mobility efforts altogether or a re-adjusted focus of policy solely on technological improvement. This would from the start preclude the chances for e-mobility to develop as a system innovation, which has to do with a general problem of new technologies (and is an argument in favor of an MLP-based ex ante analysis that looks at underlying structural change and social learning, rather than numbers): new technologies are usually expected to substitute for existing technologies fulfilling the same functions in similar ways. The frames of reference against which technologies are assessed are established use patterns, business models, market mechanisms, and cultural values, thus the social system in which technologies are embedded. This may change, because while in the beginning all the potentials, properties and new meanings of the emerging technology are still to be found out (Geels & Smit, 2000, p. 878), processes of socio-technical co-evolution can over time lead to social learning processes and a re-framing of the new technology and its social context (ibid.). It is therefore important to find a way of assessing the developments and policy initiatives in the field of e-mobility that are currently ongoing. From a sustainability point of view, such an assessment should focus on the potential of these developments to foster the emergence of e-mobility as a sustainable system innovation. As has been discussed throughout this chapter, the relevant elements of such an analysis should be social learning processes facilitating structural change and multi-level processes of socio-technical coevolution. Such an assessment is difficult and can never result in precise predictions of the future, but as shown by Kern (2012) valuable lessons about the direction of change induced by specific policy initiatives can be learned. And, as emphasized earlier in ch. 2.4, in these cases making an assessment, for each specific technology, of the likelihood that it could lead to co-evolutionary dynamics (Hoogma et al., 2002, p. 36) is important, despite the methodological difficulties this entails. Based on these considerations, in the following chapters a theoretical framework will be developed to analyze the potential, or reasons for a lack thereof, for e-mobility emerging as a system innovation. 112

126 3.3 Developing a Theoretical Framework for Analyzing System Innovations In this section, an analytical framework for studying system innovations will be developed. It will be tailored to the specific research question of this thesis regarding the potential of e- mobility emerging as a sustainable system innovation while insights may be valuable for similar cases of transitions in the making in other fields as well. It has been shown that from a sustainability perspective, the car-based transport system is faced with challenges that can hardly be met by way of purely technological solutions (such as the introduction of electric cars). Rather, a system innovation around, or including, a technological innovation such as the electric car is needed, if sustainability challenges are to be addressed in a systematic and comprehensive way. Understanding e-mobility as a system innovation thus refers to the diffusion of a new drive technology as well as fundamental changes in infrastructures, regulation, business models and user practices. How can such a complex and certainly long-term process of widespread transformation be grasped analytically? The MLP has been introduced as a framework that provides a useful heuristic for analyzing long-term and radical change. However, a number of weaknesses and difficulties in applying the MLP have been discussed in previous sections that need to be addressed, in order to approach the research question of this thesis. For instance, the macro-level perspective on transitions focusing on structural dynamics playing out over a period of 20 or more years makes it difficult to analyze transitions that are still in the making. It has been shown that in order to assess the potential of e-mobility amounting to a sustainable system innovation, it is crucial to observe and understand currently emerging patterns of socio-technical co-evolution on a micro-level. This type of analysis can shed light on the inherent potential for system innovation, the major barriers impeding structural change, and the role of actors and strategies in a concrete empirical case of such emergent processes. For developing this framework, elements will be introduced as auxiliary concepts into the broader framework of the multi-level perspective, which is a common way to utilize such heuristic frameworks in optimal ways and improve the analysis of concrete empirical cases. It has become clear that while the MLP has proven helpful in offering a comprehensive heuristic for complex transition dynamics, the downside is that due to this broad perspective, it cannot meaningfully address micro-level processes of socio-technical co-evolution at the same time. However, this is a crucial aspect when studying the potential inherent in transitions in the making. Thus, for the purpose of this thesis, some basic concepts from theory on technologybased sector transformations developed by Ulrich Dolata (2009, 2011, 2013) will be 113

127 introduced throughout the following sections, and a brief overview of this specific approach is presented in the following. In his framework for analyzing technology-based sector transformations, Dolata has developed an explicit focus on the co-evolutionary dimension in processes of socio-technical change. Basically, Dolata argues that such a framework is needed because the various approaches used to study the interrelations between technology and society focus primarily on phenomena of technological change and on how these are embedded in societal structures. What is lacking in these approaches is a focus on the way that technologies themselves shape their social surrounding and how this influences processes of technology-based transformation. It is further argued that most of these approaches remain rather abstract because their unit of analysis is an economic system or society as a whole and at this level it is hardly possible to trace the concrete patterns of co-evolutionary dynamics (Dolata, 2013, p. 25). These are in fact the specific weaknesses identified for many MLP-analyses. Dolata deals with issues of explaining the interplay of technological innovation and social change, where the focus is not merely on tracing either how technologies have changed in a social context, or on the societal framework conditions enabling technological change. He addresses the question of how technology itself shapes and changes social structures, without however adopting a deterministic perspective on technology that leaves no role for the social dimension in socio-technical transformation processes (Dolata, 2011, p. 12 f.). In order to analyze socio-technical change, Dolata develops an analytical framework that focuses on the level of concrete industrial sectors and the way that technological innovations trigger socioeconomic and institutional change processes within sectors. For instance, the emergence of advanced information and communication technologies as well as modern biotechnologies can be shown to have had an impact on various economic sectors and their political and social surroundings. In these cases, a new technology created new opportunities as well as constraints that were experienced differently in different sectors (Dolata 2009, p. 1066). Technology-based sectoral change is further explained as a function of two interrelated influencing factors (see Fig. 9). First, there is the transformative capacity of a new technology. For a specific technology, the transformative capacity may be either high, exerting direct and disruptive pressure, or low, having an indirect impact and leaving the basic sectoral structures intact (e.g. in the case of incremental innovations fulfilling subsidiary or supporting functions). Introducing transformative capacity as a distinct influencing factor allows for analyzing the impact on socio-economic and institutional structures of a technology in its own right. At the same time, transformative capacity is an inherently relational category, 114

128 because it depends not only on the specific characteristics of a technology but also on the respective sector on which it has an impact (Dolata, 2009, p f.). A second influencing factor is therefore introduced: sectoral adaptability. It describes how established sectoral structures, institutions and actors react to a new technology. Sectors characterized by low adaptability lack elements or mechanisms that allow for anticipating or proactively fostering substantial change. In these sectors, socio-technical change is disruptive and focal actors react to (rather than direct) change as an unsettling crisis. Sectors characterized by high adaptability are endowed with institutional mechanisms and actors that anticipate and actively deal with technological innovations and change. The overall sector structure is typically competitive and innovative, thus facilitating experimentation and directing structural change around a new technology in a proactive way (Dolata, 2009, p f.). Fig. 9: Technology-based sectoral change: basic categories (Dolata, 2009, p. 1067). The interplay of a technological innovation and the social processes within the sector, in which it is embedded, leads to sectoral change, which is conceptualized as a long-term process featuring phases with varying degrees of institutional and structural transformation, 115

129 ranging from path-dependent continuities to radical breaks. Looking for dynamics of gradual transformation makes it feasible to identify concrete patterns of sectoral transformation, thus tracing the complex mechanisms of stepwise institutional renewal, its interplay with constantly new opportunities and challenges presented by technological developments and the role of different actors in these processes (Dolata, 2009, p. 1073). In the following, a conceptual framework will be developed for addressing the specific research question of this thesis, i.e. how can the potential for the emergence of a sustainable system innovation in the case of e-mobility be analyzed? Based on the theoretical background discussed in previous sections and integrating auxiliary concepts from Dolata s approach of studying sector transformations, the framework developed here focuses on the following aspects: What is the relevant unit of analysis for studying system innovations? (Ch ) How can the relation and interplay between incremental innovation and radical system change be conceptualized? (Ch ) How do technological and social elements co-evolve? (Ch and 3.3.4) How to deal with the fact that the envisaged system innovation exists only potentially and in the future? (Synthesis in ch ) Unit of Analysis: Socio-technical System and Regime The MLP focuses on transitions in socio-technical systems, which are defined based on the generic societal function they fulfill. This refers to very basic functions that are socially valued, such as energy provision, mobility or housing. The boundaries of a socio-technical system are drawn in relation to such a general function. Accordingly, system boundaries can be relatively wide including all the elements and actors that are involved in the complex process of fulfilling such a particular function. This includes actors from all relevant societal domains, such as politics, business, science, civil society and the way that they shape the different system elements, such as technology and infrastructures, policy, markets, consumption patterns, culture and social practices. All those elements form a specific configuration in relation to the respective societal function, e.g. provision of energy or mobility. In terms of the MLP, there are more dominant actors within the established 116

130 configuration and more peripheral niche actors (Markard & Truffer, 2008, p. 607; Holtz et al., 2008, p. 626 ff.; Smith et al., 2005, p. 1493; Smith et al., 2010, p. 436 ff.). The meta-coordination of the various system elements can be captured in terms of a socio-technical regime. Adopting a narrow conceptualization, regimes as rule sets or institutions are clearly separated from the socio-technical system. According to this view, a socio-technical system consists of material and measurable elements, thus, obviously, technologies and infrastructures, but also for instance regulations, consumption patterns and market relations. The regime as deep structure is then a more abstract category entailing the underlying rules that make sense of the basic logic of functioning of a specific socio-technical system (Geels, 2011, p. 31; Geels & Kemp, 2012, p. 56 f.). From an analytical point of view it is important to differentiate between the regime as the deep structure, the structure or rules that actors draw upon in their concrete action and the outcome of their action which becomes manifest in the socio-technical system configuration. This analytical separation allows for tracing structuration processes even though it is clear that structure only has a virtual existence and materializes in the concrete actions and behavior of actors and eventually also the physical consequences of them (What types of technologies are produced? What types of infrastructures are being built?). However, as an interpretive approach the differentiation of structure and agency (in Giddens terms) allows for a more structured analysis and explanation of processes of reproduction as well as change in socio-technical systems. Thus, according to a narrow conceptualization, a regime consists of rule sets. Different social groups (e.g. in the spheres of science, technology, politics, markets) have different rule sets and these rules guide the behavior of actors who reproduce the elements of a sociotechnical system. In a regime various semi-coherent rule sets are linked together and mutually influence each other. In this way, a regime is understood as the deep structure that lies behind the stability and path-dependency of socio-technical systems (Geels, 2002, p. 1260; 2004b, p. 904 f.; 2006b, p. 1071; 2011, p. 27). A distinction can be made between different types of rules that play a role here. Geels distinguishes three types of rules: formal, normative and cognitive rules (2004b, p. 904). First, formal rules or institutions are the most obvious and explicit types of rules, they can be laws and regulations issued by governmental entities, or codes of conduct of other, nongovernmental organizations. They usually exist in an official or written form and they always include measures of sanctioning or punishing non-compliant behavior. Second, normative rules comprise the shared set of norms and values of a specific social group or society as a whole. They are usually implicit rules that are expressed in a more or less common 117

131 understanding of what is right or wrong, what can be expected from people, what are their roles and responsibilities. Normative rules may be characterized as a softer type of institution, in the sense that they usually do not exist in a coded form and do not include a system of formal sanctions. However, the existence of normative rules may be less explicit but is nonetheless very real. They are transported and learned via socialization processes and behavior that deviates from broadly shared normative rules will be met by sanctions in the form of irritation by other members of the relevant social group or even exclusion from it. Third, cognitive rules are the very basic belief systems, mindsets, interpretive schemes, which structure the way that actors perceive their environment, give meaning to it and process information. Cognitive rules are important for constructing meaning and making sense of experiences, thus enabling actors to interact with others and their general surrounding, while at the same time they also constrain their perception and scope of action because there is necessarily a limited amount of interpretations of the world. Cognitive rules are the most implicit type of rules or institutions; they exist largely unconsciously in peoples minds and are reflected in a social group s shared symbols or myths, their routines and common interpretive frames (Geels, 2004b, p. 904). These different types of rules or institutions can be related to Giddens modalities of structuration. The modalities are elements that actors draw upon in the reproduction of systems and thus also leading to the reproduction of structure. There are three types of modalities: interpretative schemes that actors draw upon; facility, i.e. allocative and authoritative resources that actors draw upon; and norms. These modalities provide the link between the concrete interaction in which actors engage, which can be analyzed in terms of communication, power and sanction, and the more abstract structural features, which can be analyzed in terms of signification, domination and legitimation. Giddens emphasizes that the three types of modalities can be separated only analytically and are closely interlinked in actual social practices (Giddens, 1984, p. 28 ff.). The same is true for the rules making up a socio-technical regime. The different types of rules are interlinked and together form a rule set. Such a rule set is relatively stable because the different rules build a more or less coherent whole, where they relate to each other and where changes in one rule affect others as well. Therefore, regimes can be described as semicoherent rule sets (Geels, 2004b, p. 904). Furthermore, different actor groups share their individual rule sets, which provide them with orientation and coordination. It is therefore possible to distinguish between, for instance, policy regimes, technological regimes, science regimes or consumer regimes. In a specific socio-technical system, these then form sub- 118

132 regimes that together evolve as a socio-technical regime. The socio-technical regime thus describes the meta-coordination between the different sub-regimes, which in turn can be seen as form of meta-coordination of rules in a rule system. The overall socio-technical regime usually does not encompass all the relevant sub-regimes in their entirety. Since a socio-technical system is defined in relation to a specific societal function, the socio-technical regime includes those rules that interlink the different sub-regimes in relation to that function. This explains that the individual sub-regimes making up a socio-technical regime are autonomous to some degree and at the same time also dependent on the other sub-regimes (Geels & Kemp, 2012, p. 55). The focus on socio-technical systems and underlying socio-technical regimes as the relevant unit of analysis in the MLP is a major difference as compared to similar approaches explaining socio-technical change. One example of such approaches that has been discussed here is Dolata s framework for studying technology-based sector transformation. Apart from similar theoretical assumptions regarding processes of change and similar research questions being addressed, they differ significantly with regard to the relevant unit of analysis. Since the aim here is to develop an integrated framework and refining the MLP, the unit of analysis addressed in Dolata s approach will be outlined in the following. In contrast to the decidedly systemic perspective of the MLP, Dolata argues that a sectoral approach is needed, in order to study the concrete patterns of socio-technical change and avoid overly general analyses that result from focusing on abstract economic or societal systems (Dolata, 2013, p. 25). However, the approach for studying sector transformations developed by Dolata adopts a relatively broad perspective on sectors. A sector has a relatively clearly demarcated economic core defined by the sector s main industrial or economic activity (e.g. producing cars). The way this activity is carried out, is mainly shaped by the core economic actors and structures but is at the same time also influenced by non-economic actors and institutions. A sector is therefore understood to include not only its economic core but also its societal environment, i.e. political actors, media, science and social movements, all of which have an impact on sector structures (Dolata, 2011, p. 18). Apart from a specific constellation of actors and the role of institutions, a sector is also characterized by its technological profile. The techno-structure and the institutional structure are interrelated and form a coherent match. Therefore, sectors are conceptualized as socio-technical fields that go beyond the sector s economic core and that are constituted by a technological profile, socio-economic structures and institutions, and actors and their relations and interactions (Dolata, 2011, p. 18 f., p. 33). 119

133 According to this approach, the fact that a sector is relatively stable with regard to its typical characteristics and actor constellations can be described in terms of a sectoral regulation pattern. This refers to a structuration process evolving from the interrelations between technology and socio-economic institutions: typical socio-technical constellations, structures and institutions provide a framework that guides the behavior of actors and shapes their interaction. It is important to note that in a sector, defined as a socio-technical field, there is a wide array of actors involved. There are the economic core actors, such as the established producers, service providers, suppliers and consumers, as well as more peripheral economic actors, such as developers of new technologies, and also non-economic actors, such as politicians, scientists and civil society actors. These are related in different and sectorspecific ways. Their relations and thus also the overall configuration of a sector tend to be relatively stable over long periods of time, but they are never static. The different actors have different interests and power resources, and they interpret given rules and social institutions differently. This leads to continuous incremental changes and in some instances also to radical change. The impetus for radical change often develops in the periphery of a sector and then challenges the established core actors that have long been dominating sectoral structures (Dolata, 2011, p. 29 ff.). The elements of a socio-technical field a specific technological configuration, the types of relevant actors and the role of institutions could also be part of a corresponding socio-technical system according to the MLP approach. Also, the differentiation of core and peripheral actors within a sector as a socio-technical field bears similarity with the relation of niche and regime level in the MLP. However, the decisive difference is the basic logic of delineation, which in Dolata s approach refers to the economic core of a sector. For instance, the automotive sector is defined around its economic core of producing cars (Dolata, 2011, p. 18) and it is characterized as being overall proactive and open towards new technological developments, due to technology-based competitive pressure (Dolata, 2011, p. 96). This is a valid result when considering the automotive sector per se, because there is price competition and manufacturers compete via technological innovations. However, innovations here refer to incremental innovations that remain within regime limits. From the point of view of sustainability transitions research, the car-based mobility regime is highly stable and characterized by inertia (Geels et al., 2012, p. 351). In the MLP approach, the socio-technical system with its underlying regime as the relevant unit of analysis is delineated in relation to the societal function it fulfills. Therefore, niche and regime levels are differentiated according to the degree to which established structures of fulfilling this function are being reproduced, 120

134 or whether alternative practices are being sought that might fulfill this function in a different way. The differentiation of core and peripheral actors in Dolata s sector transformation approach is in contrast based on the importance that different actors have for a very specific economic activity. Both, technology-based sector transformation and the MLP focus on socio-technical change albeit motivated differently. Dolata s sector transformation approach focuses on the impact of new technologies, how they induce change, and how change can look very differently in different sectors, thus tracing the interrelated, socio-technical nature of such change. It is recognized that this type of change can only be grasped fully when sectors are understood to be socio-technical fields, because technologies and individual firms (as core economic actors in a sector) are embedded in wider social structures. So, the empirical elements in technology-based sector transformations are similar to those relevant as elements of a socio-technical system in the MLP approach. Here, it has also been recognized that sociotechnical change is a multi-dimensional phenomenon that can only be understood when looking beyond isolated technology developments or firm strategies. The difference in perspective, however, is that the focus is on the function that a specific socio-technical constellation fulfills within a society and change of the socio-technical constellation is analyzed in relation to the respective function (Smith et al., 2005, p f.). So, Dolata s approach of sector transformation aims at analyzing the impact of new technologies on specific socio-technical configurations, while the MLP analyzes change in socio-technical configurations with a view to the overall function a socio-technical system fulfills. This explains the relevance of the MLP for studies on sustainability transitions: It is especially suitable in cases where the research question is about sustainable ways of fulfilling a societal need, such as mobility, energy and water provision or housing (Konrad et al., 2008, p. 1192). A sectoral approach is too limited for these types of questions because it focuses on the impact of a given technology on a given configuration and even though the concept of sectors as socio-technical fields is similar to a socio-technical system in the MLP, the analytical focus is on the core economic structure and activities, which is fundamentally different from the more systemic perspective on general functions fulfilled by a socio-technical system. Similarly, the technological innovation systems (TIS) approach in transition studies focuses on green innovations and conditions for their development and is thus also less suitable when interested in the overall sustainability of a socio-technical field. Choosing a suitable unit of analysis thus depends on the research question. When one is interested in the fate of a specific technology or the effects it has on specific actors and 121

135 structures, a sectoral approach is useful. When studying sustainable system innovations this focus is too narrow. This type of question is concerned with fulfilling societal needs in radically different ways and thus cannot meaningfully be limited to a specific industrial sector or set of established actors (or even a specific technology, in the sense that its characteristics and specific way of functioning and non-technological consequences are fixed in its design). In this respect, the concept of regime, which is central to the MLP, can be helpful. The fact that a regime is generally defined in relation to a generic societal function represents a central analytical strength for transition studies. It allows for introducing an analytical focus on sustainability-oriented ways of fulfilling societal functions, which is an advantage as compared to many more traditional approaches that focus on transformations around specific green technologies or the impact of green innovations on a given technological configuration. Furthermore, the regime concept is, on the one hand, relatively open in so far as it does not prescribe a specific unit of analysis; rather, it can be applied to different types of physical system configurations, depending on the respective subject of study. On the other hand, it is analytically precise in so far as it refers to the basic rule sets that explain a particular system s configuration and logic of functioning. In sum, when studying a sustainable system innovation, the suitable unit of analysis is a socio-technical system, which can be delineated along the generic societal function it fulfills (e.g. providing mobility ), and which is characterized by a specific regime, thus the rule systems or the deep structure that explain the specific way of functioning and the physical configuration of the socio-technical system. A system innovation can be defined as regime change. A major challenge in transition studies is the delineation of concrete socio-technical systems and regimes for specific cases, thus the basis for empirical operationalization. Across transition studies, conceptual papers as well as case studies, the concept of regime has been applied very differently. This has been criticized as a weakness of the MLP framework in general by several transition scholars (Holtz et al., 2008, p. 625; Markard & Truffer, 2008, p. 605; Smith et al., 2005, p. 1492). For instance, it has been argued, that the regime concept is used as a synonym for sectors or various kinds of systems without actually specifying the concrete unit of analysis (Markard & Truffer, 2008a, p. 607). This critique has been discussed by Geels (2011, p. 31), basically arguing that the incoherent application of the regime concept is not a theory-inherent problem, but rather the result of the normal difficulty of delineating the scope of an empirical case (see also chapter 3.1.4). An important aspect for delineating regimes in a clear and coherent way is to distinguish between the socio-technical system, i.e. 122

136 the involved tangible elements such as technologies, infrastructures and market transactions, and the socio-technical regime as the underlying deep structure, i.e. the formal, normative and cognitive rules. This distinction allows for utilizing the interpretive potential of the regime concept, which aims at gaining a better understanding of the structuration process where actors reproduce system elements by focusing on the underlying deep structures. This focus is inherent to the MLP and the regime concept and resulting from its theoretical background in Giddens sociological theory of the duality of agency and structure. Therefore, applying the regime concept to cases where the focus is on overall transition patterns (and not as much on structure-agency dynamics) often leads to a vague and imprecise use of the regime concept as another term for system or sector. In many cases this goes hand in hand with an inappropriate characterization of regimes as systems with an inherent intention or directional dynamic of their own, because the micro-sociological perspective on knowledgeable agents reproducing and modifying structure is lost when looking only at the macro-dynamics of transitions (ibid.). Nonetheless, delineating a regime in a concrete case is not trivial or self-evident. Drawing boundaries of a socio-technical system and defining the relevant regime for an empirical case depends on the research question. Based on the respective research question, the scope of the analysis has to be determined and then the definition of the relevant regime is the result of a deliberation process of framing the research question. For instance, in energy transition studies, the regime concept can be used to focus on centralized power generation regimes or on regimes of different energy sources, or even on regimes of different conversion technologies. In each of these cases different structures, actors and practices would play a role. This shows that defining a specific regime depends not only on the suitable aggregation level, but also on the perspective of the research interest. As a consequence, it is important to deal explicitly with the question of why and how the regime concept is applied in empirical research (Holtz et al., 2008, p. 624; Markard & Truffer, 2008a, p. 606 f.). The starting point for defining a regime is the societal function to which it is related. This focus is suitable when studying sustainability transitions or sustainable system innovations and it has implications for the definition of a regime. It spans the production, consumption and governance perspective in relation to the specific function and it looks beyond specific technologies. Therefore, the regime cannot be determined around the functioning of a specific technology or in relation to a specific policy or group of actors. All actors and elements that are relevant for fulfilling a societal function need to be considered. The regime then captures the institutions, power relations and value systems expressed in the 123

137 (re-)production of structure by actors, which materializes in the way that different actor groups interact and (re-)produce physical system elements in relation to the societal function (Holtz et al., 2008, p. 626; Konrad et al., 2008, p. 1192). Following on from the orientation along a societal function, Holtz et al. (2008) propose a definition of regimes based on five characteristics. The first characteristic of a regime is that it has a purpose, thus is delineated according to a societal function. Here, the regime is understood to be a system including the technologies, institutions and power constellations that determine the way that a societal function is being fulfilled. Second, a regime is characterized by coherence. This means that the different elements are interdependent. Technologies, infrastructures, institutions, actors and their strategies mutually adapt to each other and evolve as an interrelated and coherent whole. Third, a regime is characterized by stability. Innovations do occur, but they are incremental in the sense that they originate from established search heuristics and stay within established (technological) paradigms. As a consequence, they also do not alter established structures. In this way, the regime presents a selection environment that prevents radical innovations that might question the basic logic of a regime s functioning. More precisely, a regime is dynamically stable, allowing for incremental changes following along a trajectory with relatively clearly delimited boundaries. Fourth, a regime is characterized by non-guidance. This means that there is not one central actor steering a regime and thus the way that a societal function is fulfilled. It also means that there is no common strategy or action by all relevant actors to fulfill a specific societal function in a shared effort. Rather, the specific trajectory of a given regime is the result of the interrelated actions of different actor groups, having varying intentions and pursuing different strategies and interests. The societal function is an abstract category that provides an analytical perspective on the interrelated and interdependent actions of different actor groups in a field delineated by the researcher. The regime concept is a tool for interpreting the basic features and logic of functioning of a system, which emerges from the alignment of actors and their strategies and other system elements. Finally, a regime is characterized by autonomy in the sense that its basic characteristics and dynamics are determined by internal factors. This is not to say that a regime is not shaped or influenced by external factors, such as macro-economic or overall political trends, thus the landscape in terms of the MLP. A regime may also be affected by other socio-technical or ecological systems. However, the specific trajectory of a regime is 124

138 not mainly the result external forces, but rather develops from the alignment of regimeinternal actors and other system elements (Holtz et al., 2008, p. 626 ff.). Based on these five characteristics, Holtz et al. have adopted the following definition: A regime comprises a coherent configuration of technological, institutional, economic, social, cognitive and physical elements and actors with individual goals, values and beliefs. A regime relates to one or several particular societal functions bearing on basic human needs. The expression, shaping and meeting of needs is an emergent feature of the interaction of many actors in the regime. The specific form of the regime is dynamically stable and not prescribed by external constraints but mainly shaped and maintained through the mutual adaptation and co-evolution of its actors and elements (Holtz et al., 2008, p. 629). The definition by Holtz et al. (2008) is useful because it departs from a technology-centered perspective on regimes and utilizes the specific analytical power of the regime concept that sets it apart from other approaches. According to Markard and Truffer (2008, p. 607), the key conceptual strengths of the regime concept are that it can facilitate a better understanding of the stability of socio-technical configurations through the interlinkages between various system elements, of the many dimensions involved in fulfilling a societal function, and of innovation processes that usually produce incremental innovations that stabilize trajectories of socio-technological systems. These aspects basically focus on explanations of the dynamic stability of socio-technical systems delineated in relation to a societal function, which is also the core of the definition developed by Holtz et al. (2008). A similar approach is chosen by Konrad et al. (2008). They also define a regime in relation to a societal function and propose to identify a specific regime in cases where couplings between constituting elements (actor networks, technologies, institutions) are stronger (e.g. if they are regularly enacted or difficult to substitute) within a specific regime than outside (Konrad et al., 2008, p. 1193). This definition based on couplings also focuses on the regime-internal dynamics as an explanation for dynamic stability in the way a societal function is being fulfilled. However, in all of these definitions there is no clear differentiation between sociotechnical system (the tangible elements involved in fulfilling a societal function) and the regime (defined as the deep structure, or rule sets, underlying the physical system configuration). For instance, Holtz et al. (2008) do make reference to structuration theory, arguing that they adapt the concept of the duality of structure, in order to focus on actions 125

139 executed by the actors influenced by the structural elements and structural elements reproduced and changed by the action of actors (Holtz et al., 2008, p. 631). By including both tangible system elements and intangible structural elements in the regime concept, Holtz et al. aim at capturing structuration processes as a whole. They are in fact integrated in this way in reality, since structure has only a virtual existence (Giddens, 1984, p. 17; Sewell, 1992, p. 6) and manifests itself in the concrete action of actors and thus eventually also in the material system elements. However, it is argued here that for analytical clarity it makes sense to differentiate between agency and structure. Therefore, a narrow regime definition is adopted of regimes as the deep structure that should be viewed separately from concrete agency in socio-technical systems. This separation is only valid for interpretative purposes, arguing that [t]he crucial determining factor in the effective operation and development of the regime lies not simply in the agency of individual core members, but in the norms and procedures governing their structured relationships and interdependencies (Smith et al., 2005, p. 1505). The basic rationale of this approach is to improve the understanding of the dynamic stability in concrete configurations of socio-technical systems and their trajectories by focusing the analysis on the underlying structuration processes. In order to clearly conceptualize these processes, a separation of structure and agency as analytical categories is advocated here. The definition developed by Holtz et al. (2008) is therefore adapted in the following way: A regime comprises a coherent configuration of technological, economic and social rule systems, including formal, normative and cognitive rules, extending over different groups of actors with individual goals, values and beliefs. A regime relates to one or several particular societal functions bearing on basic human needs. The expression, shaping and meeting of needs is an emergent feature of the interaction of many actors in a socio-technical system, where interaction reflects structuration processes related to a regime. The specific form of a socio-technical system and its underlying regime is dynamically stable and not prescribed by external constraints but mainly shaped and maintained through the mutual adaptation and co-evolution of its actors and elements. Accordingly, the five characteristics of a regime are defined as structural properties. For instance, the purpose of a regime refers to the rule systems related to the way a societal function is fulfilled. These include the different rule systems of the various dimensions involved in fulfilling the societal function (technology, industry, markets, policy, culture). Coherence then captures the interdependence and mutual adaptations that have evolved across 126

140 these different rule systems forming a coherent overall regime. Stability can be interpreted in terms of the simultaneously constraining and enabling character of structure. Thus, regime structures do enable the development of innovations and at the same time have a constraining effect because they produce mostly incremental innovations. Non-guidance means that there is not one dominant rule system imposed on all relevant actors in a socio-technical system, clearly steering it in a particular direction. Non-guidance describes a situation where different rule systems, i.e. different sub-regimes, are interrelated and become aligned and how an overall regime emerges from this dynamic constellation. Autonomy means that the regime as a rule system is not merely a result of external forces. In the logic of structuration theory, actors are knowledgeable and actively create structure, they are not simply reproducing a given structure. The way that regimes are shaped can therefore not be explained as a reactive or predetermined process. There is room for creativity and alternatives for action and so regimes can be described as autonomous because their specific form evolves from the concrete actions of knowledgeable actors. To some degree, regimes are shaped by external conditions and environmental constraints; however, the concrete shape that a regime takes within such broader contexts depends significantly on the interpretations of actors. So, the stability of regimes as well as their specific development paths should be seen in a wider context, however they cannot be explained by external factors only. The development of a regime is the outcome of processes of mutual adaptation and alignment of interrelated subregimes. In sum, the starting point of the analysis of a system innovation should be an orientation along the particular societal function that is in question. This requires a careful definition, since the notion of a societal function is broad and can be ambiguous, too. As a second step, the relevant socio-technical system and underlying regime by which the societal function is fulfilled should be identified and differentiated. This is important in order to analyze structuration processes, which requires a clear analytical separation of structure, or in this case regimes, on the one hand, and agency, in this case interaction in socio-technical systems, on the other. Even though this separation is only an interpretative approach and only useful for analytical purposes (in practice such a separation does not exist), it is nonetheless decisive. The reason is that a specific physical system configuration can in principle express very different underlying structural principles or logics of functioning. If the aim is to really understand the basic logic of a socio-technical system and its trajectories, in order to identify the potential for substantial change, it is necessary to gain a profound understanding of its dominating structures and underlying guiding principles. 127

141 3.3.2 Socio-technical Change and the Concept of System Innovation A next step is then to understand how socio-technical systems and regimes change. What type of change is involved and what concrete processes play a role? The concept of system innovation needs to be defined and characterized more clearly as a process of co-evolutionary socio-technical change. According to the definition adopted in this thesis, the co-evolution of technological and social dimensions is the second major characteristic of a system innovation. Such co-evolution may be understood as the central bridge leading from technological substitution (first characteristic of a system innovation) to the emergence of new functionalities in a socio-technical system (third characteristic of a system innovation). In the MLP approach, the usage of the concept of co-evolution can be characterized as an implicit or underlying basic idea, which is expressed by focusing on socio-technical systems and regimes in the first place. As has been described earlier, a regime comprises various sub-regimes and together these form the semi-coherent deep structure of a sociotechnical system. The concept of co-evolution is used in this context to capture the complex process where each of these sub-regimes have their own specific characteristics and follow their specific development paths, while at the same time they are mutually dependent and together shape the overall trajectory and stability of a socio-technical regime. Thus, in the MLP, co-evolution describes the alignment processes and interlinkages within regimes, and dynamics across niche and regime levels (Geels & Schot, 2010, p. 21). It has been argued that a result of this analytical perspective is a lack of agency and understanding its role in multi-level transition processes (Genus & Coles, 2008, p. 1441; Smith et al., 2005, p. 1492). Even though the MLP has a view on agency, due to its roots in Giddens structuration theory and thus the basic understanding that structure or regimes as rule sets are always enacted by specific actors (Geels, 2011, p. 29), it produces explanations that are generally of a more structural nature (Foxon, 2010, p. 8). This relatively broad focus on the interplay of structuration processes within and across the three levels does indeed abstract from micro-level dynamics, such as the development of specific technologies, strategies of individual actors or actor groups and how such processes are interrelated. However, this critique refers to the application of the MLP to specific case studies and is not an inherent shortcoming of this approach in general. The way that co-evolution is understood in the MLP allows for analyzing the dynamics within and between different rule systems of the various technological and social dimensions involved in a socio-technical system and forming a specific alignment and trajectory. It is therefore important to consider that regimes are not homogeneous entities but rather dynamically stable configurations of 128

142 semi-coherent rule sets. This means that there is neither complete inertia within a regime, nor can specific technologies be assumed as givens because they are shaped by structural dynamics and at the same time can themselves have varying impacts, i.e. it is not fixed in their design how they shape their environment in detail. Geels argues accordingly that such a detailed understanding of regimes would make the strength, homogeneity and internal alignment of regimes an empirical question rather than an assumption (Geels, 2011, p. 31). Even though a regime may appear as a coherent and homogeneous unit from the outside, and even though it is possible to determine basic structural principles of a regime as a whole, a second and more rigorous glance will show that there are almost always internal dynamics and tensions. Such internal dynamics are the inevitable consequence of the fact that a regime comprises various sub-regimes. There are typically technological regimes as well as policy, science, market and cultural regimes, each characterized by their own structural dynamics. It seems quite likely that not all of the rules constituting these different rule systems match each other and that there will be tensions, expressed in debates or conflicting interests of the different actor groups. The overall regime remains stable when there are basic rules or guiding principles prevailing across the different sub-regimes and when they are aligned in such a way that dynamics in one sub-regime are curbed by the interlinkages with other subregimes. However, there may be situations where tensions or deviating dynamics between specific types of rules or even between entire sub-regimes de-stabilize a regime as a whole. At this point, a window of opportunity emerges for a substantial change of the regime, and thus a transition (Geels, 2011, p. 31; Geels & Schot, 2010, p. 44). Due to the complexity of regime internal dynamics, radical changes are likely to proceed through a complex and unique history of interrelated events (Smith et al., 2005, p. 1494). With regard to the type of change depicted in the MLP, a transition or a system innovation is not a linear process, where clear causal relationships between a limited number of influencing variables can be identified. In that way, it deviates from classical concepts of the diffusion of innovations, conceptualizing change as a sequence of typical phases of an S- shaped curve (cf. Rogers, 1962). The MLP also is not a technology-centered approach, even though technologies are a central element of the focal unit of analysis, namely socio-technical systems. The emphasis of the MLP is on structural conditions, because change is perceived to be a systemic and social phenomenon and not mainly the result of technology-push. While classical innovation approaches focus on the development and diffusion of specific technologies in a wider societal context, the MLP focuses explicitly on this context, because 129

143 systemic change (a transition or system innovation) cannot be explained adequately by limiting the analysis to the emergence of a new technology (Geels & Schot, 2010, p. 27). The concept of co-evolution can be seen to work twofold in the MLP. First, from an overall perspective it is the co-evolution of the three levels of the MLP niches, regime and landscape that may lead to a transition in a socio-technical system. A transition is defined as a change of the regime, which may come about when an established regime is weakened, for instance due to landscape pressure, and a window of opportunity emerges for a niche innovation to diffuse and become dominant, which then leads to a new regime developing around this niche innovation. Radical change is thus conceptualized as the result of coevolution, in so far as the dynamics within and between the three levels are interlinked and eventually reinforce each other amounting to a full-fledged transition of the socio-technical system. Second, the concept of co-evolution plays an important role in the conceptualization of regimes. Since a transition or a system innovation is defined as regime change, it is important to understand the processes within a regime. It comprises different sub-regimes and the way that these sub-regimes are mutually dependent and together develop within a relatively stable trajectory is explained as a co-evolutionary process. Since the sub-regimes are interlinked and evolve together, they either stabilize a certain development path, or if diverging tendencies in one sub-regime become strong enough, they may mutually reinforce overall regime de-stabilization. Even though the MLP is often characterized as a co-evolutionary approach, the closer look adopted here shows that the explanations offered by the MLP to deal with radical change may more accurately be described in terms of interlinkages and alignments within and between the levels of a socio-technical system (Foxon, 2010, p. 8). It may thus be argued that in the MLP a constructivist explanation of technology s generation on the local level is combined with a social evolutionary approach of structural selection on the global level (Geels & Schot, 2012, p. 30). In sum, when trying to characterize the type of change depicted by the MLP and the way that the concept of co-evolution is understood and applied, it becomes clear that the MLP is a heuristic analytical tool with a focus on the macro-level processes of transitions. At the micro-level, the development and diffusion of technology is viewed as an outcome of multilevel structuration processes. These structuration processes are of central interest and their complex interrelatedness is described in terms of co-evolution. However, there is no explicit focus on concrete co-evolutionary dynamics at a micro-level, for instance between specific elements of a socio-technical system, between specific groups of actors or between specific 130

144 sub-regimes. This is due to the heuristic character of the MLP, which aims at offering a comprehensive picture of the long-term and dynamic macro developments of transitions. In order to make the MLP fruitful for empirical studies and concrete cases, it should be complemented with other theoretical approaches and a suitable scale and scope need to be chosen. This is also the reason why in this thesis the notion of system innovation (even though often used synonymously) is preferred to the notion of transition, in order to be able to focus also on the local or micro-level of more direct and concrete socio-technical co-evolution in socio-technical systems. With regard to the overall research question of this thesis, the point is not to focus on innovation processes as regards the way that technologies are being developed and how this development is shaped by wider socio-economic processes, because the electric car as a technological innovation does already exist and its diffusion is supported in different ways. However, the central question is whether it has the potential to develop as a radical and systemic innovation thus, looking at how this technology shapes its socio-economic environment and understanding how its usage or role is in turn shaped in the context of a broader socio-technical system. Thus, there is a need to understand and conceptualize coevolutionary change, where co-evolution means something in between the MLP s sociological/structural co-evolution of aligned sub-systems and more classical (co- )evolutionary approaches focusing on variation, retention and selection mechanisms in innovation processes. The latter understanding of co-evolution has its origins in biology and refers to the way that two populations, species or organisms evolve and mutually influence each other. Concepts of evolution and co-evolution have, however, also been applied in the social sciences, focusing on organizations, institutions or technologies as (co-)evolving systems. Important categories are variation, retention and selection, which are the basic processes of evolution, explaining how novelties emerge (mutation of genes in the biological context, or the development of innovations in social systems), how advantageous characteristics of a population or a social system are retained or preserved over the long term, and how the process of selecting favorable characteristics works ( by accident in the biological world or in an often partly guided way in social systems, such as markets or political arenas) (Foxon, 2010, p. 12). One may speak of co-evolution when there are two evolving systems, i.e. characterized by processes of variation, retention and selection, which have a direct and causal impact on each other. The type of impact they have on each other may vary, because co-evolutionary relationships can be anything from cooperative to predatory in character. 131

145 Kallis and Norgaard explicitly emphasize that [c]oevolution is different than mere codynamic change although they are often misused synonymously (Kallis & Norgaard, 2010, p. 690). Concepts of evolution and co-evolution have been adapted from their origins in biology to the field of Science and Technology Studies, in order to explain technological change. Technologies or technological systems have been conceptualized as evolving systems, explaining their often path-dependent development characterized by specific technological trajectories or paradigms (as discussed in ch and 3.1.2). It has also been recognized that technological development or change cannot be adequately understood when observed in isolation, i.e. as a purely technological phenomenon. The concept of co-evolution is used to understand how technologies are shaped by their socio-economic environments and how, in turn, technologies shape economic and other social systems. Each of these systems is viewed as an evolving system and the way that they influence each other, thus their coevolution, leads to dynamic and uncertain technological development paths. It is thus also emphasized that no specific technological development is the inevitable outcome of a particular technology-inherent logic. There is thus also rarely the possibility to predict sociotechnical developments and there are always potential alternatives (Foxon, 2010, p. 8; Geels & Schot, 2010, p. 32). However, it has already been pointed out that the concept of co-evolution, in its many adaptations in the field of the social sciences, is not used in a coherent way and, what is more, is often used imprecisely. Especially in transition studies, what is termed co-evolution can in many cases be described much more accurately as co-dynamic development. The basic idea here is generally to show that socio-technical transitions can be understood only from an overall systems perspective, and that various technical and social elements are interlinked and mutually influence each other. This may be characterized as an overall co-evolutionary perspective, but it falls short of tracing the concrete causal links between two evolving systems and identifying the concrete mechanisms and dynamics of such co-evolutionary processes. In order to improve the analytical rigor of transition studies, the understanding that everything might be coevolving with everything else needs to be complemented with the identification of what is coevolving with what and how in specific conditions or contexts and as relevant to specific analytical and policy purposes (Kallis & Norgaard, 2010, p. 691). Analytically separating the particular elements that are co-evolving (against the background of a systemic perspective, where all those elements are aligned in various ways) may improve understanding of the concrete micro-level processes relevant in transition processes and thus 132

146 may also cast new light on the roles of actors, strategies or policies in these processes (Foxon, 2010, p. 13). Compared to the MLP, Dolata s concept of technology-based sector transformation proves to be of greater strength with regard to this aspect. In this approach, the transformative capacity of new technologies and sectoral adaptability are analytically separated, in order to focus on the direct interrelation between technological and social elements. New technologies are shown to have a transformative capacity and thus a direct impact on a sector s socioeconomic and institutional structure. At the same time, the degree to which a sector is adaptable determines the actual impact a new technology can have. So, co-evolution is conceptualized here as a direct interrelation between changes in technology and changes in other elements of a sector. Before focusing on this interplay between the transformative capacity of new technologies and sectoral adaptability in more detail, the type of change and basic dynamics that Dolata s sector transformation approach entails will be discussed in the following. It has already been noted in ch that in contrast to the MLP, the focus is on change in industrial or economic sectors. Nonetheless, sectors are defined rather broadly as socio-technical fields, including not only the economic core of a sector but also other socio-economic structures, institutions and actors that are related to it. In that way, the research questions and basic problem frames of the MLP and of the sector transformation approach differ substantially, while their empirical object of study can be relatively similar. This applies not only to the elements of a sociotechnical field as defined by Dolata or a socio-technical system in the MLP, but also to the basic dynamics within these fields or systems and thus also the basic understanding of fundamental change within them. For instance, similar to the conceptualization of socio-technical systems and regimes, sectors as socio-technical fields are not to be seen as completely homogeneous entities made up of actors that share the same set of rules and apply the same strategies in aiming to achieve a certain goal in unison. A sector is rather characterized by the interplay of heterogeneous groups of actors with different interests and different degrees of power to influence others. This results in continuous processes of structuration, of de- and re-institutionalization that eventually shape the configuration of a specific sector. The overall outcome may be more or less coherent when looking at a specific sector as a whole, but underneath this there are power struggles, conflicts and differing interpretations (Dolata, 2011, p. 26). It is argued that fundamental change in a sector is often triggered by the emergence of new technologies. The potential of new technologies is in many cases hampered by the 133

147 established institutional structures of their socio-economic context, which then act as a barrier to the further development of new technologies and inhibits reaping their potential benefits. In this way, new technologies can trigger processes of searching for new arrangements within organizations and established institutions because they cast doubt on whether established socio-economic structures are still fulfilling specific functions adequately. Since sectors as socio-technical fields consist of heterogeneous actors, those search processes or early phases of change are characterized by a high degree of uncertainty. There are different interpretations of the problem at hand, different strategies being pursued and different goals or orientations being referred to by different actors. These different points of view are competing in this phase and there are power struggles and open conflict between different actor groups. It can also be observed that these search processes are often initiated by relatively marginal actors that are part of the socio-technical field but do not belong to its economic core. The result of search processes developing around new technologies may in some cases result in fundamental changes in a sector, i.e. changes in its basic socio-economic structure, its institutions and organizational arrangements. This means that there may in the end be new actor configurations and new power constellations, new patterns of market relations, new demands on the relationship with science or on policies and regulation as well as new boundaries being drawn of a specific sector and re-adjusting its relations with other sectors (Dolata, 2011, p. 26 f.). The type of change in this concept is similar to the conceptualization in the MLP. Both approaches view change as a gradual and multidimensional process that unfolds in a heterogeneous and dynamically stable system. Both refer to structuration processes and view change as a non-deterministic, open-ended process characterized by uncertainty and conflicting dynamics. Fundamental change is conceptualized as a deep and structural type of change at the level of institutions, rules and guiding principles of a socio-technical system or a sector as a socio-technical field, respectively. Both approaches also emphasize the role of marginal or new actors in initiating change niche actors in the MLP and actors at the fringes of a sector in Dolata s approach. Also, both approaches refer to the concept of co-evolution, whereas it is applied differently. In the MLP, co-evolution is used as a notion that describes the alignment and interlinked development of the different dimensions of a socio-technical system across the levels of niche, regime and landscape. Co-evolution also captures the coordination between the different sub-regimes making up an overall socio-technical regime. Thus, co-evolution in the MLP could rather be understood as the co-dynamic development characterizing macro- 134

148 level transition processes. In the sector transformation approach, co-evolution is applied on a micro-level, focusing on the direct impact technology has on socio-economic sector structures and vice versa. Sector transformation is explained as a function of two interrelated influencing factors, namely the transformative capacity of a new technology on the one hand, and of sectoral adaptability on the other. By analytically separating technology and the socioeconomic system of a sector, this approach aims at tracing the direct and causal links between these two influencing factors and thus explaining the resulting socio-technical change on a micro-level. Dolata explicitly criticizes the MLP for a lack of analytical rigor and accuracy at the micro-level. He argues that the basic idea of the MLP is useful, since it depicts sociotechnical change as a process where pressure for change is exerted on regimes as selection environments and where this pressure is perceived and acted upon in specific ways. The interplay of selection pressure and the capacity to adapt may then result in a transition. However, he criticizes that it is not explained sufficiently how selection pressures emerge and what exactly determines the degree of adaptive capacity of a regime. Furthermore, the types of regime change discussed in transition studies are not systematically developed and derived from theoretical insights on the interplay of selection pressures and adaptive capacities. Therefore, he argues, that many transition studies are limited to providing plausible explanations that lack the necessary theoretical rigor and empirical basis (Dolata, 2013, p. 27; Dolata, 2011, p. 40). Consequently, in his approach Dolata focuses on the direct causal links between the transformative capacity of new technologies and sectoral adaptability. However, even though technology-based sector transformation provides an approach that allows for producing explanations on the direct and causal influences of technologies on socio-economic sector structures and vice versa, it still falls short in terms of a truly coevolutionary explanation. Neither technology, nor the sector is conceptualized as an evolving system per se. This aspect is much more explicit in the MLP. The different dimensions or regime elements in a socio-technical system (technology, markets, policy, culture) are conceptualized as rule systems that follow trajectories and their development can be explained by taking recourse to evolutionary mechanisms of variation, retention and selection. What is then missing is a micro-level focus on co-evolution, i.e. the causal impacts that two or more of those evolving sub-systems directly have on each other, as developed in Dolata s sector transformation approach. With regard to the research question of this thesis, Dolata s approach is problematic because it is a sectoral approach. When one is interested in the fate of a specific technology or 135

149 the effects it has on specific actors and structures, a sectoral approach is useful. However, as has already been pointed out in ch , when studying sustainable system innovations this focus is too narrow. This type of question is concerned with fulfilling societal needs in radically different ways and thus cannot meaningfully be limited to a specific industrial sector or set of established actors. Therefore, when studying a sustainable system innovation, the suitable unit of analysis is a socio-technical system, which can be delineated along the generic societal function it fulfills (e.g. providing mobility ), and which is characterized by a specific regime, thus the rule systems or the deep structure that explain the specific way of functioning and the physical configuration of the socio-technical system. Therefore, the type of change that is in the focus is a system innovation defined as regime change. An important aspect of a system innovation is socio-technical co-evolution and it has been shown that in the MLP, the various social and technological dimensions or sub-regimes are well described in terms of evolving systems. With the notion of co-evolution, the MLP rather describes co-dynamic alignment processes capturing the bigger picture of mutually dependent developments across niche, regime and landscape levels. What is lacking is a clear view on the direct and causal influences that social and technological elements have on each other, i.e. actual co-evolutionary mechanisms on a micro-level. Such a perspective is necessary for dealing with cases such as e-mobility, where a technology is emerging and it is important to understand what its impact is on social system elements and how concrete coevolutionary dynamics shape the way the technology is used in (potentially) new system configurations. If a transition or a system innovation will develop around such a technology, the configuration of the resulting socio-technical system and the way it fulfills a specific function, can be explained by the very concrete and direct interactions of a socio-technical coevolution process. In order to be able to adapt such a perspective, the concepts of transformative capacity of new technologies and sector adaptability will be adapted, where necessary, and integrated as complementary concepts within the wider MLP-framework. The analytical separation of a new technology and of the relevant social elements will help to better trace direct and micro-level co-evolutionary processes. It is argued that such an integration is possible because basic premises of the two approaches, MLP and sector transformation, are similar. Their general view on change as a gradual and complex process corresponds, as well as the way that socio-technical systems or socio-technical fields are conceptualized as dynamically stable configurations characterized by structuration processes. A major difference, however, is the central boundary criterion for delineating the relevant unit of analysis. In the MLP it is the societal function that a socio-technical system can be said to 136

150 fulfill, while Dolata focuses on sectors as socio-technical fields defined around an economic core. Since the focus of this thesis is on sustainable system innovations, which requires a view on fulfilling societal functions in socio-technical systems, the next step in developing a theoretical framework is to adapt the concepts of transformative capacity of new technologies and of sectoral adaptability in such a way that they can be incorporated in an MLP-framework and still retain their analytical strength The Transformative Capacity of New Technologies In order to analyze and better understand technology-based transformations of industrial or economic sectors, Dolata introduces the concept of transformative capacity of new technologies as one decisive influencing factor. The point is to conceptualize technology as an independent influencing factor as a first step, in order to be able then to grasp its interplay with the social elements, eventually resulting in sectoral transformation. Nonetheless, it is clear that this separation is analytical in nature and that the transformative capacity of new technologies is an inherently relational concept. It can only be determined based on the possibilities to use and further develop a new technology in a specific sectoral context and based on the institutional and structural conditions which a new technology meets in a specific sector. Technologies can have an impact on sectors only because applying or using them usually requires some degree of change in socio-economic institutional structures. It is the mis-match between the potential of a new technology and the feasible possibilities of using it under given circumstances which can be characterized as the transformative capacity inherent in a new technology. The more a new technology becomes relevant for a sector and the less it can be embedded in existing sectoral structures, the larger is its transformative capacity and thus the pressure for change it exerts on established sectoral organizations, institutions and behavioral patterns (Dolata, 2011, p. 68 f.). Considering these qualifications regarding the transformative capacity of new technologies as a relational category, while at the same time analytically isolating technology as an influencing factor in its own right, two general characteristics of a new technology are introduced by Dolata. First, it has to be distinguished whether a technology can be characterized as exogenous or endogenous, i.e. is it developed within the sector that is to be studied, or outside? This is an important differentiation because it allows for locating where the pressure for change is coming from. It can make a difference whether a new technology has been developed within a sector and thus has already been shaped by sectoral structures in its early development, or whether it comes from outside and has been developed under 137

151 substantially different circumstances. In both cases, the transformative capacity can be either high or low. A technology may require enormous structural adaptations even though it has been developed within a sector. At the same time, the transformative capacity of an exogenous technology can be low because it may only have a subsidiary function not affecting the technological core of a sector. So, determining whether a technology can be characterized as endogenous or exogenous is helpful for understanding where the pressure for change originates but it says nothing about the degree or strength of this pressure (Dolata, 2009, p f.). Second, it has to be distinguished what type of effect a new technology has on a specific sector. Effects may range from being only indirect or even supportive of established sectoral functionalities to being much more direct and disruptive with regard to the established structures of a sector. Based on these effects, a new technology can be characterized as having a low or high transformative capacity, respectively (Dolata, 2009, p. 1068). So, in order to get a grip on the transformative capacity of a new technology, it should be considered where it originates, i.e. whether potential pressure for change is exerted from within a sector or from outside its boundaries, and it should be assessed whether it has a direct and disruptive impact on established actors, institutions and structures of a sector or whether its effects are more indirect and complementary, thus leaving the established sector configuration largely intact (Dolata, 2009, p. 1069). Based on these two characteristics it can then be determined whether the transformative capacity of a specific new technology is either high or low in the context of a specific sector that is confronted with the emergence of this technology. It should be noted explicitly at this point that transformative capacity cannot be measured quantitatively in any meaningful way. It is hardly possible to identify unambiguous and universally valid indicators for such an undertaking. This is partly due to the fact that transformative capacity remains a relational category and that it cannot be determined based on characteristics or features of a specific technology in its own right and independent of its sectoral context of application. However, a number of qualitative criteria have been identified according to which the transformative capacity of a new technology can be captured in a way that is precise enough to serve as a useful heuristic for developing scenarios or reconstructing specific cases. These criteria assess the characteristics of a new technology with regard to the specific configuration of the sector that is to be studied (Dolata, 2011, p. 70 f.). 138

152 The criteria, as presented in the following, aim at grasping the extent to which new technological opportunities and their attributed peculiarities alter the technological profile of the sector and enhance or destroy the existing knowledge base and competencies; affect the existing patterns of research and development, production, distribution, products, and market relations; facilitate or enforce new patterns of cooperative and competitive interaction; initiate institutional readjustments (e.g. new regulative and legal frameworks or modified guiding principles, norms, and beliefs); open up or widen the existing borders of the sector, thereby provoking a more intense interpenetration of different sectors (e.g. by the migration of powerful external actors into the sector) (Dolata, 2009, p. 1070). The basic idea of the concept of transformative capacity as reflected in these criteria is that a new technology may have the potential to shed new light on the established mechanisms and basic logic of functioning of a given sector. The emergence of a new technology may give rise to a situation where the existing technological basis of a sector, its institutional and organizational structure as well as its guiding principles are increasingly being questioned. Put differently, new technologies may raise doubts about an established and to date unquestioned sectoral configuration because they challenge the existing match between technology, structures, and institutions in the course of their formation and adoption (Dolata, 2009, p. 1070). In such a situation, a new technology causes pressure for change and it may trigger processes of searching for new sectoral configurations. However, the fact that a new technology possesses a high transformative capacity does not necessarily have consequences in practice. Whether or not the emergence of such a technology leads to processes of sectoral transformation depends significantly on the way that this is perceived and acted upon by sectoral actors. Transformative capacity merely points to the inherent potential of a technology in relation to given sectoral structures, thus as a concept it captures the virtual range of possibilities or alternatives that become feasible with this technology. Actual change results from the interplay that unfolds within the social configuration and dynamics of a sector (Dolata, 2011, p. 73 f.). In order to study system innovations, the concept of transformative capacity, which has been developed in the context of sectoral studies, needs to be adapted to the context of socio-technical systems. It has already been pointed out that according to Dolata sectors can 139

153 be defined as socio-technical fields around an economic core. The elements included in such a socio-technical field are similar to those in socio-technical systems as the focal unit of analysis in the MLP. The major difference is that socio-technical systems in the MLP are delineated according to the societal function they fulfill. Therefore, the relevant economic core activity is not as central as in the sector transformation approach; rather, it is one element making up the socio-technical system in concert with politics, technology, infrastructure and culture. Following this, the criteria for assessing the transformative capacity of a new technology as developed by Dolata need to be slightly adapted, in order to fit with the perspective on the function fulfilled by a socio-technical system. Again, since the definition of a sector as a socio-technical field is similar to that of a socio-technical system, most of the relevant aspects are included in the existing criteria. Thus, following the original systematization by Dolata: In order to assess the transformative capacity of new technologies it needs to be analyzed to what extent the new technological opportunities and their attributed peculiarities alter the technological profile of the socio-technical system and enhance or destroy the existing knowledge base and competencies; affect the existing patterns with regard to market relations, research and development, policy making and user behavior; facilitate or enforce new patterns of interaction between different actor groups of the socio-technical system; initiate institutional readjustments (e.g. new regulative and legal frameworks as well as modified guiding principles, norms, and beliefs within different sub-regimes); open up or widen the existing borders of the socio-technical system. Adapting the criteria in this way is possible because the basic idea they reflect of the concept of transformative capacity is that of a technology emerging as a challenge to existing structures and pointing towards alternative ways of organizing these established structures. This idea is central both for understanding co-evolutionary processes of change in sectors as socio-technical fields as well as in function-oriented socio-technical systems. The difference in perspective when focusing on the function fulfilled in a socio-technical system has been introduced in the criteria by including, apart from the economic dimension, the political and cultural or user dimension more explicitly (these elements also play a role in the sector 140

154 transformation approach but are viewed as a field forming around the economic core). So, even though the criteria can be kept similar and the basic idea of the approach remains the same, the empirical consequences and the outcomes with regard to the types of answers may differ more fundamentally. The change in perspective from focusing on a sector and its economic core activity to a socio-technical system that fulfills a societal function in a specific problem field (e.g. sustainable mobility) may not alter the focus being set on a specific technology (e.g. electric cars) and it may in both cases be useful to apply the concept of transformative capacity, in order to understand the challenges and potentials inherent in this technology. However, with regard to the implications that this type of research can have, it is a major difference whether the transformative capacity of electric cars is analyzed in relation to a sector (possibly changing the way that cars are produced) or in relation to a functionoriented socio-technical system (possibly changing the way that mobility is provided). Assessing the transformative capacity of a new technology along the criteria developed here can be shortly summarized as an approach that identifies the mis-matches of a new technology with the established elements of a socio-technical system (e.g. its technological profile, patterns of interaction, or even its boundaries) and thus also with the underlying regime (e.g. the guiding principles within different sub-regimes). Since all the physical elements and interactions of actors within a socio-technical system can be understood as an expression of the dominant regime, it is important to trace how a new technology affects this underlying deep structure. Empirically, the focus obviously has to be on the tangible system elements, in order to trace the implications of a new technology. However, a regime perspective will allow for a more systematic analysis of the most important and fundamental mis-matches related to the very basic logic of functioning of an established system. Another empirical consequence is a research approach that focuses on the technological shortcomings that are attributed to a new technology. Technological shortcomings understood as mis-fits of a new technology with an established system are, in principle, shortcomings only with respect to the existing conditions and thus may in some cases very clearly point to the new technology s transformative capacity System Adaptability In Dolata s technology-based sector transformation approach, sectoral change is explained as a result of the interplay between the transformative capacity of new technologies and the adaptability of the sector. The concept of sectoral adaptability is introduced as an analytical tool to capture the way that newly emerging technological opportunities are perceived and 141

155 dealt with in a given sector. Adaptability in this context means more than a passive reaction to changing conditions, as in the common sense of the word. Since new technologies are seldom emerging as a given with clear implications on ensuing consequences, adaptability here refers to a more active way of dealing with the uncertainties and ambiguities involved in technological change processes. New technologies often remain in an immature and developing phase for quite a while. At this early point they do cause pressure for change on sectors because it is already clear that there are mismatches with existing structures. However, it is much less clear how exactly a new technology can and will be used in the future, how significant its role will be and thus, how sectoral structures will change exactly. As a result, new technologies first and foremost trigger search and selection processes in a sector, where different expectations are formulated, discussed, revised and refined in an often controversial and non-linear fashion. The concept of sectoral adaptability captures the way that a sector is able to deal with such search and selection processes in situations where the established organizational and institutional structures seem to function less well than they did before and where well-tried strategies and guiding principles fail to deliver solutions to new problems (Dolata, 2009, p. 1070; Dolata, 2011, p. 76). In order to assess the adaptability of a sector, it is not sufficient to focus on the way that already mature technologies have eventually been adopted. At this point the critical phases of sectoral change have already been overcome. Sectoral adaptability can only be grasped when looking at the process as a whole and especially focusing on the early phases of (potential) change. Whether a sector is characterized by high or low adaptability depends on how the potential of immature technologies as well as the related organizational and institutional effects it might have are perceived and dealt with. Thus, sectoral adaptability describes the degree to which a sector is prepared to deal with insecurities and ambiguities concerning its future development (Dolata, 2011, p. 76 f.). Dolata identifies three levels on which sectoral adaptability can be identified: on the organizational level the ability to identify, communicate, and adopt the challenges of new technologies in a timely manner and to renew established routines and strategies; on the institutional level the openness and flexibility to change and readjust [ ] the regulations, norms and shared beliefs guiding and structuring the activities of the actors involved; 142

156 on the structural level the permeability of research, production, market, and demand conditions in supporting discontinuous innovations, developing new products, constituting new markets, and facilitating the entry of new actors (Dolata, 2009, p. 1070). Based on this more general separation of levels of a sector, Dolata identifies a number of factors that are relevant for determining the adaptability of a sector. These are (1) the ability of key actors to adapt as well as the institutional and structural conditions for overall sectoral adaptation; (2) the role of the fringes of the sector and the relations between the sectoral core and its periphery; (3) organizational, institutional and structural mechanisms fostering or impeding change; and (4) power relations within the sector based on social or technological sources of power (Dolata, 2011, p. 118 ff.). Characterizing a sector according to these criteria and analyzing how these factors are interrelated in the specific sector allows for identifying its specific degree of sectoral adaptability. Generally, this may range from low to high adaptability. Low adaptability can be found in sectors that have been stable and successful over long periods of time without having had to adjust substantially their market relations, industry structures, regulations and overall guiding principles along the way. These sectors are often highly vulnerable with regard to technological changes and at the same time are lacking mechanisms or anticipatory strategies for dealing with uncertainties and prospective change. The emergence of new technologies is mainly perceived as a risk threatening established and successful economic performance. As a consequence, these sectors typically react to the emergence of potentially disruptive new technologies by increasing their efforts of preserving the established sectoral structures and technological profile. A transformation of these types of sectors usually is a crisis-ridden process where the impetus for change comes from actors at the fringes of a sector or from external actors. These actors perceive the emergence of new technologies as an opportunity to gain importance and possibly even replace the established core actors, who are trying to preserve the existing structure to their advantage. So, essentially processes of change in these cases can often be characterized as power games (Dolata, 2009, p. 1071). In contrast, there are sectors characterized by high adaptability, in the sense that they are more anticipative of novel technological developments and actively deal with radical technological innovations. This is possible in cases where a sector possesses established institutionalized mechanisms for anticipating and adapting to changing circumstances or new 143

157 influencing factors. Usually this can be found in sectors that are characterized by fierce competition and where innovation and experimentation are important and highly-valued elements. Such general conditions allow for a proactive attitude of actors and structural flexibility to deal with emerging technologies already at an early stage (Dolata, 2009, p. 1071). As regards the criteria that have been developed for assessing the transformative capacity of new technologies, it is obvious that also sectoral adaptability cannot be approached by any clearly delineated quantitative measure. The aim of the criteria for sectoral adaptability developed by Dolata is to provide a heuristic to assess sectoral adaptability in a systematic way, while being aware of the fact that there may be a multitude of relevant factors and interlinked dynamics playing a role. Furthermore, this approach is a way to deal with the fact that causes for inertia and structural path dependency have been relatively well explored, whereas high adaptability and its causes are much less clear (Dolata, 2009, p. 1072). This may in fact simply be due to the nature of the matter at hand, namely that sectors as well as sociotechnical systems generally develop in such a way that stability and continuity can be preserved as much as possible. The MLP departs from exactly this insight and provides a way of addressing the core analytical puzzle of transitions, namely stability and change in socio-technical systems (Geels & Kemp, 2012, p. 50). Overall system stability is explained as a result of pathdependent development and lock-ins of different types, for instance shared beliefs that make actors blind for developments outside their scope; consumer lifestyles, regulations and laws that create market entry barriers; sunk investments in machines, people and infrastructure; resistance from vested interests; low costs because of economies of scale (Geels & Kemp, 2012, p. 50 f.). These mechanisms correspond to the factors that Dolata discusses as an explanation of low sectoral adaptability, whereas the focus of the MLP is broader, in order to capture the function-oriented socio-technical system, which goes beyond Dolata s focus on the core economic activity determining a sectoral socio-technical field. Also, similar to Dolata s approach, where external actors and actors at the fringes of a sector play an important role with regard to introducing radical novelties, the MLP emphasizes the role of niche-actors 5. They are important advocates of change, since established actors and structures The discussion of the role of fringe or niche actors points to the importance of entrepreneurs and an overall entrepreneurial culture. Fostering the potential of emerging system innovations in the field of mobility as well as other crucial societal sub-systems therefore depends on more profound approaches of entrepreneurship education 144

158 are threatened by innovations that are radical in so far as they imply a different system configuration (Geels & Kemp, 2012, p. 50 f.) Rather than systematizing varying degrees of adaptability, in the MLP it is assumed that socio-technical systems are in principle characterized by overall dynamic stability. Nonetheless, fundamental change is possible and has occurred throughout history. Such fundamental changes, i.e. transitions or system innovations, are conceptualized as regime change. A recent approach introduced by Kemp et al. (2012) for studying ongoing or even just potential future transitions, is to investigate cracks (p. 3) in established regimes. It is argued that regimes as the deep structure of socio-technical systems shape and stabilize pathdependent trajectories, while at the same time there are always regime-internal dynamics due to the fact that a regime consists of various sub-regimes. A potential for substantial change is emerging when tensions are developing within the regime, i.e. when cracks in the thus far well-aligned regime configuration threaten overall stability. Such a situation presents a window of opportunity for radical innovations to diffuse and trigger wider socio-technical change (Geels et al., 2012, p. 355 f.; Nykvist & Whitmarsh, 2008, p. 1385). Similar to Dolata s concept of sectoral adaptability, in the MLP it is the appearance of cracks in the regime that increase the possibility for a more proactive perception and dealing with novelty, a larger scope for institutional readjustment and thus eventually the prospect for a transition or system innovation. In order to improve the analysis of system innovations in terms of co-evolutionary socio-technical change, it is proposed here to adapt the concept of sectoral adaptability to the study of socio-technical systems. Since both the MLP and the sector transformation approach emphasize the importance of structuration processes, where technological innovations may trigger new practices and institutional readjustments, it is possible to integrate the concept of adaptability in the MLP. Since the focus of the MLP is on socio-technical systems, the notion of system adaptability is introduced and will be spelled out more clearly in the following. The aim is to develop criteria of system adaptability (similar to what Dolata has done for the study of sector transformations), in order to enrich the MLP with a more systematic approach for analyzing co-evolutionary processes of socio-technical change. In the MLP framework, a system innovation can be defined as regime change. Therefore, system adaptability, analogous to the concept of sectoral adaptability, describes the in schools and universities but also in professional contexts in practice and in individual companies. These issues are elaborated in: Braukmann et al. (2012), Braukmann & Schneider (2008). 145

159 way that a technological innovation is perceived and dealt with in the context of established regimes. Since regimes are defined as rule sets, system adaptability depends on the stability or instability of the regime and thus the way that rule systems can be readjusted. Regimes are conceptualized as rule systems based on Giddens theory of structuration and the basic idea of the duality of structure. In the context of socio-technical systems this means that a material change in such a system is always interrelated with changes in structures or, more accurately, it is the outcome of a structuration process where actors (re-)produce structures. In that sense, system adaptability increases, when a regime as the established, and so far persistently reproduced, set of rules is destabilized. Since both the sector transformation approach and the MLP assume structuration processes as the basis for the development of specific system configurations and their stability, a closer look should be taken at how change can be explained by drawing on Giddens theory of structuration. The basic idea of the duality of structure is implicit in the sector transformation approach and more explicitly used in the MLP in distinguishing different levels according to their varying degree of structuration. However, even in the MLP the reference to Giddens concept remains relatively superficial. Stability and change are interpreted in terms of actors reproducing established structures or actively introducing novelty by deviating from established structures. This does result in a useful heuristic for understanding the scope of action that different actor groups perceive for themselves and make use of. However, what is needed is a more in-depth understanding and an actual explanation of how structuration processes may produce change (see also ch ). This may then shed light on the concrete factors influencing system adaptability. In this respect, it is important to reflect on the underlying question that Giddens structuration theory deals with. The concept of the duality of structure explains how large parts of everyday social practices are characterized by routinized behavior and how this contributes to the overall stability and cohesion of social systems. However, this does not mean that structuration theory first and foremost explains social order or the continuous reproduction of the very same social practices over and over. Rather, structuration theory is about stability in the sense that social practices are conceptualized as a continuous phenomenon, while within these continued social practices there is room for change as regards the concrete shape and form of these practices (Schiller-Merkens, 2008, p. 180). This is reflected in the MLP in so far as regimes are described as dynamically stable. Based on this, the question is how can a change in structure, i.e. in institutions or basic rule sets, be explained? 146

160 As a starting point for approaching this question, it is important to note that Giddens distinguishes three dimensions of the duality of structure, along which different types of institutions can be grouped. These three dimensions are signification, domination and legitimation and their specific properties are reflected in different types of interaction in social systems, namely in communication, in the way that power is exercised and sanctions are enforced. Giddens then introduces the modalities of structuration, in order to clarify what the specific types of institutions are that actors draw upon in their social practices and in what way this leads to the (re)production of overall structural properties. The modalities of structuration include (1) interpretative schemes that actors draw upon when they communicate and which reflect structural properties with regard to signification, (2) facility, i.e. the allocative and authoritative resources decisive for the execution of power and reflecting structures of domination, and (3) norms which imply different types of sanctions and reflect structures of legitimation (Giddens, 1984, p. 28 ff.). Institutions, or rule sets in terms of the MLP, can be grouped along these modalities. Institutionalized forms of symbolic meaning and discourses can be analyzed in terms of interpretative schemes, political and economic institutions reflect the allocation of resources (and thus power relations) and moral or legal institutions express what is perceived to be legitimate behavior and are connected with sanctions. Grouping and systematizing institutions in this way allows for analyzing and explaining the specific configuration of a social system, for instance by identifying the role of different types of resources or by studying the way that symbolic meaning or legitimized traditions have an impact on social practices and what types of sanctions they produce (Lamla, 2003, p. 58 f.). Based on this general conceptualization, possibilities for structural change can be identified when there are tensions or contradictions emerging in the way that the modalities of structuration are drawn upon. The conditions for structural change are thus favorable when there is heterogeneity or even contradictions among the rules in a social system. Since rules according to Giddens refer to structures of signification and legitimation this means that within a social system interpretive schemata and expectations regarding appropriate behavior would vary. With regard to authoritative and allocative resources, the conditions for structural change are favorable when the distribution of these resources is unequal and there are large disparities in the distribution of power (Schiller-Merkens, 2008, p. 198). The analysis of system adaptability should thus focus on the interrelations and dynamics between the different sub-regimes, i.e. the interlinked rule sets that together constitute an overall regime. The key elements determining whether overall system 147

161 adaptability is high or low are reflected in regime configurations that are either well-aligned or characterized by internal tensions. Much like transitions and regime change in general, system adaptability can ultimately be determined only in retrospect. However, it is possible to identify landscape factors that have an impact on established regimes as well as dynamics between niches and a dominant regime. This latter aspect of niche-regime dynamics is important because even if there is substantial landscape pressure and a niche with transformative capacity, the potential for a transition depends on the way that incumbent regime actors deal with this situation. In some cases, cracks in the regime may induce regime incumbents to take up niche-innovations. This can play out favorably, because regime actors are in a more powerful position than niche actors and have the resources to introduce novelty at a larger scale (Nykvist & Whitmarsh, 2008, p. 1385). An example from the field of mobility could in this respect be the recent efforts of large OEMs to develop carsharing services, thus promoting and developing this niche. However, studies of transition management policies have also shown that in many cases incumbent regime actors manage to capture green niches, such as for instance carsharing, and sustainability discourses, in order to fit them into current regime structures as a way to preserve the status quo (Avelino, 2009, p. 381; Meadowcroft, 2009, p. 336; Voß et al., 2009, p. 289). It is thus important to identify not only the cracks in the regime, but also the way that these are dealt with, i.e. how adaptable the socio-technical system actually is. The key elements determining whether overall system adaptability is high or low are reflected in regime configurations that are either well-aligned or characterized by internal tensions: Rules: What are the dominant interpretive schemes and norms in a socio-technical system? Are they coherent across the regime as a whole or can tensions and conflicts be identified? Resources: How are resources distributed in a socio-technical system? Can conflicts or tensions with regard to the distribution of power be identified? Assessing system adaptability should thus focus on constellations of power between actors and the guiding principles or frames of reference they draw on. System adaptability increases with emerging conflicts regarding the distribution of power among groups of actors and incoherence and emerging uncertainty with regard to established norms and values. 148

162 3.3.5 An Integrated Analytical Framework: Synthesis and Discussion Integrating elements from Dolata s concept of technology-based sector transformation in the Multi-Level Perspective is possible because both approaches emanate from a similar conceptualization of change in general. Both go beyond more traditional and simpler approaches of looking for alternating phases of system inertia interrupted by short periods of radical system shift. Transitions as well as sector transformations are conceptualized as continuous and ongoing processes, determined by the constant interplay of stability and varying degrees of change. The basic aim of both approaches is to identify concrete patterns that may help structuring and better understanding overall complexity. By integrating their particular strengths in an adapted analytical framework, research on sustainability-oriented system innovations, such as e-mobility, can be structured in a more profound and theoretically grounded way (cf. also Augenstein, 2014 forthcoming, 2012a, 2012b). The different elements that have been discussed in the preceding sections the relevant unit of analysis, conceptualizations of socio-technical co-evolution and system innovation will now be integrated in an adapted analytical framework (see Fig. 10). The core of the analytical framework is the system innovation, defined as regime change. Since the regime refers to the deep structure that lies beneath the physical, i.e. measurable, configuration of the socio-technical system elements, the analysis will empirically have to focus on those elements, such as technologies, infrastructure, formal and informal institutions as well as actor strategies and behavior. This is how the two influencing factors from the concept of technology-based sector transformation the transformative capacity of new technologies and sectoral adaptability can be integrated. They are assessed according to a number of qualitative criteria and at different levels, respectively (Dolata, 2009). These are corresponding to the measurable elements of socio-technical systems according to the MLP. However, their analysis in terms of factors influencing a system innovation requires a broader focus transcending sector and current regime boundaries, investigating the potential for radical system shift. Therefore, it is referred to system adaptability, instead of sectoral adaptability, because a system innovation implies that a societal function may eventually be fulfilled in substantially different ways, with a different technology and possibly also by different actors. The analytical separation of transformative capacity of new technologies and of system adaptability allows for tracing concrete patterns of co-evolutionary transformation. Dolata has criticized the MLP approach for identifying transition patterns and pathways that have been derived in an interpretative fashion that is not adequately systematic and thus 149

163 lacking validity. At the same time, Dolata s approach remains limited, due to its sectoral focus at least when system innovations are concerned. Thus, integrating an analytical separation of transformative capacity of new technologies and system adaptability, on the one hand, and the MLP concept of a wider socio-technical system for fulfilling a specific societal function as the relevant unit of analysis can improve the analysis of system innovations. Fig. 10: An integrated framework for analyzing system innovations. Introducing the concepts of transformative capacity and system adaptability can shed light on the co-evolutionary dynamics emerging between a specific technology and relevant actors. It is argued that this can add to MLP analyses by focusing on the role and strategies of concrete actors in dealing with technological innovation, and by offering a perspective on technology as an influencing factor in its own right. Tracing the case-specific interplay of a technology with specific characteristics and actors in more or less stable system configurations provides a way of assessing the potential emergence of a system innovation. The potential for a system innovation thus depends on the interplay of the transformative capacity of a new technology and the adaptability of the socio-technical system in question. In each specific case, it needs to be determined whether the transformative capacity and system adaptability are (relatively) 150

164 high or low, respectively. In principle, four types of situations are possible regarding this interplay, as shown in Tbl. 3. This table shows that in the case of a new technology characterized by low transformative capacity, diffusion is feasible no matter how adaptable the socio-technical system is. This is obviously the case, since low transformative capacity means that the technology can be characterized as an incremental innovation that fits within the current structures and simple market-based diffusion is likely. However, in this situation there is no potential or need for a system innovation, or at least none that builds around the technology in question. In the case of a new technology characterized by high transformative capacity, the potential for a system innovation depends on the degree to which the socio-technical system is adaptable. Where system adaptability is low, even the diffusion of the new technology, let alone the emergence of new functionalities, is not feasible at all. This describes the ideal-typical situation where a technology does not fit the existing structures, which in turn are stable and hardly able to adapt. If, however, system adaptability is high, there is a high potential for a system innovation to emerge, because in order to deal with the mis-fits of the new technology, the system necessarily has to adapt (in terms of developing new functionalities). Tbl. 3: The potential for system innovation: transformative capacity and system adaptability. This type of classification can help to clarify a few aspects in the debate about the momentum of the developments around the electrification of the car and it helps to shed light on the question whether there is a potential for sustainable system innovation. A simple diffusion process of electric vehicles seems unlikely, because it can be assumed that the transformative capacity of electric vehicles is high, in so far as they do not fit the current mobility system and cannot compete with conventional cars on the established system s terms (e.g. regarding range and price). If the transformative capacity of the electric vehicle (which will be assessed in more detail in ch. 4.3) is to unfold, thus if its system-innovative potential 151

165 can be realized, this will crucially depend on the degree to which current mobility systems can be characterized as adaptable, i.e. in terms of regime cracks and the specific ways these are dealt with by actors in concrete situations. And, finally, even if patterns of socio-technical coevolution and signs of the emergence of new functionalities can be observed, this does not necessarily mean that a new system configuration is automatically more sustainable than the current system. Thus, criteria for assessing the contribution of emerging developments or trends to more sustainable mobility patterns need to be added to the analysis. Referring back to the three traps for sustainable system innovation, it can be shown that the quantitative and the qualitative trap are closely interlinked when it comes to radical innovations with high transformative capacity: a quantitative diffusion of electric vehicles will not be sufficient to address complex sustainability challenges, and at the same time qualitative change in terms of new functionalities is not only relevant from a sustainability perspective, but also if a radical innovation that does not fit seamlessly in an established system is to diffuse at all. In short, if e-mobility is to be developed successfully and if it can be shown that electric vehicles have a high transformative capacity, then this is only feasible in terms of a system innovation. While such a process of socio-technical co-evolution is a precondition for developing more sustainable system configurations, this is not automatically the case and not every kind of new functionality is necessarily producing more sustainable outcomes. Especially this latter aspect leads to a brief reflection on the strengths and also the limitations of the framework developed here: In sum, this integrated framework for analyzing system innovations aims at emphasizing the specific strengths of the MLP and addresses its weaknesses by refining it based on insights from the concept of technology-based sector transformation: focusing on socio-technical systems and regimes as the relevant unit of analysis when interested in sustainable system innovations, and a clear conceptualization of socio-technical co-evolution in the context of sustainable system innovations. It will not be feasible, with the help of this framework, to determine with certainty whether e-mobility will eventually emerge as a sustainable system innovation and what type of transition pathway will unfold. As already discussed in earlier chapters, a transition or system innovation can only be identified in retrospect and the dynamics in transitions are characterized by uncertainty and contingency. However, the framework allows for an assessment of potential by characterizing the potential inherent in a new technology, in this case the electric vehicle, determined by its transformative capacity, the potential inherent in current system dynamics, i.e. its adaptability, and at least early signs of co-evolutionary patterns. If the basic uncertainty and contingency involved in transition processes are taken serious, and considering the 152

166 history of the electric vehicles and its past failures, early signs of emerging system-innovative trends should be treated with caution. A focus not only on positive signs but just as much on the barriers and obstacles for system innovation seems appropriate because the collective research focus on change has tended to obscure a potentially more important reality, both in theoretical and empirical terms there is not enough attention paid to understanding, confronting and ultimately resolving the tendency for change not to happen (Wells & Nieuwenhuis, 2012, p. 1682). In the following empirical analysis, the framework developed here will be used to identify episodes where system-innovative dynamics begin to emerge, but its explanatory value will be more pronounced when it comes to assessing the barriers and obstacles impeding their further development. 153

167 4 E-MOBILITY AS A SUSTAINABLE SYSTEM INNOVATION THE CASE OF GERMANY When interested in the potential for further momentum of developments in e-mobility, Germany is an interesting case at the moment because it has launched one of the largest national initiatives for fostering e-mobility. The aim of Pt. III is to explore the theoretical framework and the concepts of transformative capacity and system adaptability by tracing the dynamics of the German innovation system for e-mobility, focusing especially on the automotive state of Baden-Württemberg. Insights will be gained regarding the conditions for increasing system adaptability or barriers impeding it as a way of estimating the potential of this particular niche and its chances for future up-scaling. The results of the analysis are directly relevant for the empirical case, due to the critical situation of Germany with a nation-wide energy transition project underway and an important automotive sector, and with e-mobility linking these two fields that are of major relevance economically and with regard to sustainable development goals. The results will also contribute to a better understanding of transition processes that are in the making and that require conceptual approaches that are able to deal with ex ante assessments of the potential for emerging system innovations. Advancing theory in this field of transition studies is relevant for a number of fields, including but not limited to energy and transport, where technologies are central, but co-evolutionary socio-technical transformation will be imperative when aiming at more sustainable societies or societal subsystems. 4.1 Methodology The empirical analysis of the development of e-mobility in Germany will be structured along the main theoretical concepts developed in chapter 3, thus system innovation, transformative capacity and system adaptability. The aim is to trace how the transformative capacity of the BEV is adapted to, or fail to be adapted by the socio-technical system in which it emerges and thus, indicating the potential for e-mobility to develop as a sustainable system innovation. The first step is therefore to capture the abstract notion of a sustainable system innovation for the concrete empirical case of e-mobility. Developing an idea of what this would entail in practice, is an important framework for guiding the overall analysis. This will be done in chapter 4.2. Since it is not clear whether a system innovation will occur at all and even if so, what the future socio-technical system would look like specifically, the approach chosen here is to define a basic ideal-type. Defining such a basic ideal-type is sufficient for 154

168 this specific analysis, because the aim is not to conduct a full-fledged sustainability assessment of different forms of e-mobility, or a life-cycle assessment of the environmental impact certain vehicle technologies have. Rather, at this early stage of development the goal is to gain a robust idea of the direction in which things are moving. Since the potential transition is an open-ended and long-term process, a more precise approach is simply not feasible, because a sustainable mobility system may take different forms with regard to infrastructure, business models or behavioral patterns. At the same time, it is imperative to include some form of sustainability assessment to ensure that what is actually analyzed is the potential for achieving some variant of sustainable e-mobility. Therefore, literatures on sustainable mobility in general are drawn upon to identify the relevant basic criteria, such as renewable energy sources for powering vehicles, integrated and intermodal mobility patterns that do not solely rely on privately owned cars. Such elements of an envisioned sustainable mobility system of the future and, based on the theoretical framework, the concept of regime change that these elements amount to, provide an ideal-typical orientation for the analysis. Consequences for the concrete empirical design are, for instance, that basic criteria of sustainable forms of e-mobility will be deducted that serve as an orientation for identifying concrete examples, e.g. e-mobility projects with small-scale system-innovative potential, in the subsequent case study. Second, the transformative capacity of the BEV will be discussed in chapter 4.3. Based on the criteria developed in the theoretical framework, it will be analyzed in what ways the BEV (1) alters the technological profile of the socio-technical system; (2) affects established patterns within the socio-technical system (market relations, R&D, policy making, user behavior); (3) facilitates new patterns of interaction between actors; (4) initiates institutional readjustments; (5) opens up or widens existing borders of the socio-technical system. The transformative capacity will be analyzed based on a review of the relevant literatures. These include basic handbooks on the BEV or e-mobility as a new technology, studies on the potential and challenges of e-mobility in Germany that focus on technological and market-related aspects. This approach relying on secondary sources is feasible because the electric vehicle has been invented already more than 100 years ago and its basic functioning is well-known. The implications of this technology and its new role in visions of future mobility are widely discussed in the context of the German innovation system and, more generally, in public discourses on transport policy and developments in the automotive sector. Since these debates are dominated by established regime structures (rules, norms, guiding principles involved in the way that mobility is perceived, organized and envisioned), 155

169 they offer a good insight in the technological shortcomings identified in relation to this regime, thus further highlighting the basic tenets characterizing this specific regime and simultaneously offering a glimpse at the transformative capacity inherent in the new technology (whether or not this materializes eventually, or not). Third, the concept of system adaptability will be explored in a specific case within the German innovation system: the region of Baden-Württemberg. The rationale for choosing to do a case study is that this method is most suitable for studying a contemporary phenomenon in depth and within its real-life context, especially when the boundaries between phenomenon and context are not clearly evident (Yin, 2009, p. 18). This type of complexity is characteristic of most social phenomena, where multiple variables with different degrees of influence and their specific interrelations play a role. Causal explanations are thus rarely at the same time necessary and sufficient, but rather depending on complex causal chains and contingent dynamics (Mayntz, 2009, p.21). The case of e-mobility as an evolving phenomenon of socio-technical co-evolution is no exception in this respect. In addition to overall complexity in a real-world setting that cannot be manipulated like a laboratory experiment and fuzzy boundaries between the object of study and its context, case studies are also suitable when the research question and design are exploratory in nature (Baxter & Jack, 2008, p. 545). This criterion applies as well, since the aim is to explore the potential of e- mobility as a sustainable system innovation in a concrete situation, where neither the empirical basis nor the existing literature provide enough substance for identifying all relevant variables or testing pre-formulated hypotheses. Nonetheless, it is emphasized that basic theoretical propositions that guide the case study analysis are important. The concepts developed in the theoretical framework will serve as such guide for the analysis, as is laid out in this section. With regard to the concrete case study design, dealing with complexity requires multiple sources of evidence, which are then converged by means of triangulation (Yin, 2009, p. 18). The detailed design of the case study on Baden-Württemberg, including the mix of sources and methods of analysis, will be explained in chapter Baden-Württemberg has been chosen as a single case study, because it is a critical most likely case, which is one rationale for choosing a single case study design, even though this is often viewed as not sufficient for generalizing from empirical findings. It can however be argued that a single case study is ideal for generalizing using the type of test that Karl Popper called falsification (Flyvbjerg, 2006, p. 227). 156

170 Popper himself used the now famous example all swans are white and proposed that just one observation of a single black swan would falsify this proposition and in this way have general significance and stimulate further investigations and theory building. The case study is well suited for identifying black swans because of its in-depth approach: What appears to be white often turns out on closer examination to be black. (Flyvbjerg, 2006, p. 228) What does this imply for the case of e-mobility in Germany and Baden-Württemberg specifically? Based on the theoretical framework, it can be assumed that the interplay of a technology with transformative capacity and a system characterized by tensions or crisis, thus to at least some extent adaptable, would result in a system innovation, or at least the potential for a system innovation should be observable at an early stage, as is the case here. Overall, existing literature and public debate rather indicate that the prospects for e-mobility developing as a sustainable form of mobility are bleak. So, the question is, what is the actual potential for a system innovation? The case of Baden-Württemberg is suitable to explore this, because it is a critical case, which can in general be defined as having strategic importance in relation to the general problem (Flyvbjerg, 2006, p. 229). For one thing, the case of Baden-Württemberg as an innovation system for developing e-mobility is not a typical small-scale niche. It has been initiated by the federal government, is well-funded and involves incumbent actors this may imply possibilities for change at a larger scale, or illustrate why change is impeded from the beginning. It also shows characteristics that according to the theoretical framework may be signs of relatively high system adaptability: an automotive industry that is acting on global markets, thus facing global trends that turn out to threaten established industry structures and market dynamics; the participation of incumbents from the automotive industry, the energy sector and ICT companies, thus a situation where actors from different regimes cooperate and it can be assumed that tensions between different rule systems, power struggles regarding the allocation of resources and a process of reflecting on and refining different frames of reference may emerge. This makes it a critical case, selected carefully based on these characteristics, which is preferable to an average or random case here. It is also a most likely case: If it can be shown that even in this case, exhibiting relatively advantageous conditions, the potential for a system innovation remains limited, it can be generalized that this is valid also for most other cases (Flyvbjerg, 2006, p. 230 f.). Thus, the rationale for choosing a single case is that, with respect to generalization of findings, one case is sufficient to falsify a theoretical proposition. However, the basic aim is 157

171 not to simply reject the theoretical framework that has been developed in its entirety, but to refine general propositions with regard to the emergence of system innovations or transitions. Also, a flat-out rejection is not feasible, because the theoretical framework developed here is a relatively broad heuristic, which does not contain explanations of linear causality between concrete variables and thus, the case study aims at identifying the critical factors and can aid the development of more concrete hypotheses. The case is used as an in-depth exploration, analyzing where exactly the barriers for system innovation can be found and how developments are hampered. Apart from refinements of theory, this can provide insights into what may be drivers for system innovation, or at least starting points for improving the prospects for sustainability transitions in the field of transport. The three chapters on system innovation (ch. 4.2), transformative capacity (ch. 4.3) and the case study focusing on system adaptability (ch. 4.4) are not isolated, much like the theoretical concepts on which they are built represent an analytical separation rather than isolated elements that are in fact clearly distinguishable in practice, and therefore there are several interlinkages between the chapters. For instance, the ideal-type definition of a system innovation will guide the analysis by providing selection criteria for e-mobility projects that the case study is built around (the detailed case study design is layed out in chapter 4.4.1) and it will be the basis for a critical reflection of the case study findings. Furthermore, since transformative capacity and system adaptability are interrelated categories, which only have meaning when considered together, the respective findings have to be integrated as well. The findings of chapter 4.3 show to what degree the BEV can be said to have an inherent transformative capacity in the face of current regime structures. For a more detailed assessment, in chapter 4.4 the case study provides a more in-depth perspective on how this technology is actually dealt with in a specific innovation system by different groups of actors. Thus, chapters 4.3 and 4.4 are basically a shift in perspective, looking at the phenomenon of socio-technical co-evolution from both sides, the technical in chapter 4.3 and the socio- in chapter 4.4. This analytical separation will not be clear-cut throughout the entire analysis, because the two elements are co-evolving; however, the emphasis is different depending on which side of the coin we look at. In chapter 4.5, the results of the empirical section will be drawn together and reflected upon in an integrated fashion. Implications for the case of e- mobility in Germany and for the theoretical framework will be discussed. 158

172 4.2 Delineating the System Innovation as an Ideal-Type In order to deal with the question of whether co-evolutionary dynamics and the emergence of new functionalities actually contribute to a more sustainable system configuration, a sustainability assessment has to be included. Due to the complexity of the issue, this can only be done in a qualitative fashion and depends on the specific case. Transitions to sustainability are characterized by uncertainty and ambivalence, i.e. one cannot determine with certainty whether ongoing dynamics will eventually result in a transition and one can rarely define unambiguous sustainability criteria especially in cases of transitions in the making (Konrad et al., 2008, p. 1192). It is therefore proposed here that when analyzing ex-ante cases, such as the electrification of the car, a first step would be to delineate the envisaged system innovation for the specific case. Since it is not clear whether a system innovation will occur at all and even if so, what the future socio-technical system would look like specifically, the approach chosen here is to define a basic ideal-type. In order to do so, the concept of e-mobility as a system innovation needs to be connected with general criteria of sustainable mobility. Departing from current sustainability challenges, some basic criteria a sustainable mobility system would have to fulfill can be defined. It has been argued in chapter 2.2 that sustainable mobility requires a paradigmatic change including not only efficiency increases in vehicle technology and fuels but also an overall reduction in the volume of travel, reduction of trip lengths and substitution of individual car use by intermodal transport and carsharing. Following from this, some basic criteria for assessing developments around the electrification of the car can be defined. Criteria for sustainable forms of e-mobility include: more integrated and intermodal forms of mobility, concepts of collective ownership of cars or carsharing, and renewable energies as the basic electricity source. It is of course less clear what exactly the corresponding socio-technical system would look like, because there may be different types of infrastructure, business models or behavioral patterns that could make up such a sustainable mobility system. However, the concept of regime change in this case a break with the currently dominant car-centered outlook on mobility and the basic criteria required for sustainable mobility provide an ideal-typical orientation for the analysis. The aim of the following sub-sections ( ) is to link these general principles of sustainable mobility with e-mobility. This is important because in current discourses it is 159

173 often taken for granted that e-mobility, in terms of the electrification of the car, is automatically contributing to a reduction of the environmental burden caused by the transport sector. However, this causality is not as straightforward when considering that there is a global increase in the volume of traffic and thus an increase in absolute levels of resource use in production of vehicles and emissions, especially when considering current electricity mixes. Thus, technological improvement and efficiency gains will be compensated for by the sheer force of increasing absolute numbers. Approaching e-mobility as a project of technological substitution becomes even more inadequate when looked at from a broad sustainability perspective, including issues of social justice in terms of access to mobility and livable cities. The literature on sustainable mobility shows that the dominance of the car as the central technological artifact and the lock-in that has evolved around it, in terms of infrastructures, business models, use patterns and symbolic meaning, are at the heart of current problems. A transition towards more sustainable forms of mobility would thus include that the role of the car in society is at least questioned and possibly even redefined. The central question of this thesis is whether developments in the field of e-mobility link up with such more basic debates on modern transport systems and may even contribute to a situation where it becomes possible to question fundamental and long-established truths in mobility. Therefore, ideal-typical from a broadly defined sustainability perspective elements of sustainable forms of e-mobility are outlined in the following. It is shown how new linkages are beginning to show between e-mobility and the broader context of an energy transition to renewable energies (ch ), e-mobility and less car-dependent forms of mobility in fleet applications (ch ) and intermodal transport concepts (ch ), as well as between e- mobility and innovative business models in the field of carsharing (ch ). These aspects have been chosen because they are potentially contributing to more sustainable forms of mobility in principle (reducing and substituting for conventional car use) and they show the specific potential of e-mobility going beyond technological improvement. They will later on serve as indicators or basic criteria for identifying system-innovative e-mobility projects and initiatives in the case study (ch ). The purely technological issues and efficiency-related benefits in the process of electrifying the car are not dealt with in detail, since they are a minor aspect in the context of a broader sustainability perspective. Also, the reduction of the overall volume of travel and trip lengths is not considered explicitly, because this has mainly to do with substituting mobility with virtual forms of communication and with city and land use planning. These can 160

174 be considered as framework conditions within which mobility patterns unfold and thus fall outside the scope of this thesis E-mobility and Renewable Energy The first characteristic of sustainable forms of e-mobility seems to be relatively straightforward and obvious. If electrifying transport is to contribute to reducing negative environmental effects, it is not enough to reduce local emissions, which is already the case when electric vehicles are used, even in the context of an energy system where a large share of generated electricity is produced from fossil sources. A truly sustainable form of e-mobility should guarantee that the electricity used to power electric vehicles is produced from renewable sources. The specific way that the energy and the mobility transitions are interlinked and may create synergies for achieving more sustainable outcomes in both fields has been laid out in chapter 2.3 and it has also been highlighted that there are some basic challenges. This is important when looking at the envisaged transition towards sustainable e- mobility in practice. Even though there is such a large potential of electric vehicles fostering the transition to renewable energies by providing short-term storing capacity on the one hand, and the energy transition fostering sustainable forms of e-mobility by providing clean energy sources on the other, ways still have to be found for realizing this potential in practice. With current use patterns of cars and current business models of energy utilities this will not be feasible. For instance, from the point of view of energy utilities it is important that a significant number of electric vehicles is available at regular times and with a reliable storage capacity, which is difficult to organize in a system of predominantly private car ownership (Canzler & Knie, 2011, p. 113 f.). For car owners, on the other hand, it is important to have an adequately developed recharging infrastructure, which however is too expensive for energy utilities to set up, as long as it is just another sales channel for electricity (ibid.). One possible starting point for dealing with these issues is to focus on vehicle fleets (operated by companies, municipalities or other public bodies) rather than private car owners, at least at this early stage of developing e-mobility E-mobility in Fleets Fleet operators are used to managing relatively large numbers of vehicles in the most efficient way. They also have a history with electric vehicles, because early delivery services, e.g. postal or dairy products, were once preferably operated with electric vehicles. Even today, many services, delivery or other, offered by fleet operators typically remain within distances 161

175 that would be within the range of an electric vehicle. Furthermore, integrating controlled charging systems into the normal business of a fleet operator seams feasible and it would provide them with a new source of income via offering battery reserves in addition to mobility services. This would allow for a more reliable and predictable use of electric vehicles as storage capacity and even as part of bidirectional vehicle-to-grid (V2G) concepts (Canzler & Knie, 2011, p. 50; Dijk et al., 2013, p. 139). Fleets could also be an important starting point for introducing significant numbers of electric vehicles. Since in fleets the maximum capacity of vehicles is utilized and they are used in standard ways and similar use patterns, learning effects with regard to their optimal use can be realized relatively quickly. Once the fleet is put to optimal use, the electric vehicles can demonstrate their specific qualities. Furthermore, cost savings and scale effects can be reached relatively quickly in fleets, so that in this case, as opposed to the case of private car owners, both the fleet operators and the energy utilities would more easily find business models that are beneficial to both parties (Hein et al., 2013, p. 1655; Vallée et al., 2013, p. 63). For instance, in this case the charging infrastructure could more realistically be not just a charging spot but just as much an infrastructural element for guaranteeing grid stability thus, an actual connecting hub from which both sides can profit. This could also allow for developing business models that are more profitable than building up charging points for merely selling electricity (Canzler & Knie, 2011, p. 67). The fact that the prospect for electric vehicles used in fleets is much better than for normal private cars, does not necessarily mean that they will forever be confined to this particular niche. From a sustainability perspective it would be beneficial if more mobility uses or trips were organized by vehicles operated in fleets (thus reducing private car ownership and use). Especially with the electric vehicle that can be utilized as a means of transport and an infrastructural element in the energy systems, new options for business models may emerge that could trigger new, profitable forms of carsharing and intermodal mobility concepts (Canzler & Knie, 2011, p. 46 f.; Tran et al., 2012, p. 332) E-mobility and Intermodality Since sustainable mobility in general would require a reduction and substitution of car use with a shift to other modes, such as public transport, walking or cycling, sustainable e- mobility would ideally combine the introduction of electric vehicles with the development of intermodal transport concepts. E-mobility can in this case be understood as a holistic transport concept including various (electrified) modes of transport that are interlinked in a mobility 162

176 service scheme. Especially in cities there is a large potential for such concepts, because many younger people already rely on a combination of different modes of transport and do not own a car (Newman & Kenworthy, 2011). Problems of congestion and lacking parking spaces can also be addressed by a shift to intermodal mobility in cities. Furthermore, if electric vehicles are integrated in intermodal transport chains their disadvantages as compared to conventional cars, especially concerning range, do not play a significant role anymore. Even though forms of intermodal mobility can already be observed in practice, it is still a challenge to further establish marketable concepts that can compete with so far dominant modes of private car use for most trips. Successful concepts need to be reliable, easily accessible and intuitively usable, in order to be an attractive alternative to the privately owned car (Canzler & Knie, 2011, p. 38). Ideally, an intermodal mobility service would include the existing public transport system (trams, metros, local trains etc.) as well as carsharing services, rental cars, taxis and rental systems for bikes and pedelecs. All of these transport modes should be integrated in a comprehensive infrastructure of central transfer points between modes, charging and parking stations. In order to guarantee user-friendliness, a guidance and information system needs to be in place as well as a ticketing system that allows for one-stop-shop solutions, i.e. buying tickets for an entire trip using different combinations of transport modes. Therefore, a central mobility service provider is needed that can bundle and integrate activities and serves as contact point for users (Kemp & Rotmans, 2005, p. 34; Vallée et al., 2013, p. 75). A seemingly obvious type of actor that could function as a comprehensive mobility service provider are public transport operators. They already have ticketing services in place and operate mobility chains, combining for instance trams and busses on different lines. However, in Germany, public transport operators are providers of public service tasks, whereas strategically planning and financing these service tasks is in the hands of public authorities. This structure hampers entrepreneurial activities on the part of public transport operators and denies them the legal and financial means to integrate bike and car rental or sharing concepts in their services (Canzler & Knie, 2011, p. 82 f.). Similar to the situation regarding the integration of e-mobility in the energy system, here the synergies that the development of e-mobility and intermodal transport concepts could produce are so far not really utilized because the existing actor structure, business models and use patterns impede innovative approaches. 163

177 4.2.4 E-mobility and Carsharing As already mentioned, carsharing can become a central element of an intermodal transport system. However, it will be treated here also as a separate element of sustainable e-mobility because it represents a very specific form of reducing the burden of transport that cannot be substituted by a non-travel activity. In this case, an environmental benefit does not result from a reduction of the volume of transport or trips lengths, or from a shift to other modes, it rather increases efficiency by increasing the average occupancy of existing (shared) cars and it questions the deeply embedded and dominant pattern of the privately owned car as the most common means of transport (Bongardt & Wilke, 2008, p. 66; cf. Wilke, 2009a, 2009b). Even if carsharing does not represent a shift towards fundamentally different, e.g. intermodal, forms of transport and rather a tendency to move away from car ownership to car services, there can still be substantial sustainability benefits. In a service-oriented scheme it is likely that the symbolic value of the car, which is central when it is owned and in this way connected to an individual personality, becomes less relevant and criteria of efficiency and operability gain importance (Nykvist & Whitmarsh, 2008, p. 1382). In Germany, the earliest carsharing systems have been established already in 1988 and today a number of different forms of carsharing exist. There are smaller-scale concepts where privately owned cars are shared among a group of people, e.g. in a neighborhood, or can be rented via a private online exchange platform. More common are carsharing services offered by professional providers. The various professional services differ with regard to the range of vehicle types that is on offer, the modalities of registration as a user and of booking a car, the billing system (paying either by kilometers driven or by time of use), and whether or not the cars have to be returned to specific pooling stations (Vallée et al., 2013, p. 72 f.). Carsharing systems have increasingly emerged over the past few years, with also large automotive companies (Daimler and car2go ) and public transport actors (Deutsche Bahn and DB Flinkster ) founding their own services. A reason is that especially younger people living in cities do not buy as many cars as they used to or buy them at an older age and that, in general, European markets for cars are increasingly saturated (Infas, 2011). Realizing that many people do in fact rely on a combination of transport modes, an opportunity is seen to profit from these newly emerging social practices by offering short-time rental services that fit within intermodal mobility patterns (Canzler & Knie, 2011, p. 59). This general trend can play an important role for the successful introduction of electric vehicles, and it contributes to more sustainable forms of transport. A major obstacle for the diffusion of electric vehicles on mass markets is that battery costs are so high that they cannot 164

178 compete with the cost of a conventional car. If electric vehicles are operated as shared cars, users are not faced with the high initial purchase costs and carsharing operators can much better realize cost and scale effects via their fleet management systems and by operating their BEVs also as means of electricity storage which might be an additional source of income, for instance through feed-in tariffs or reduced electricity prices (Dijk et al., 2013, p. 139; Vallée et al., 2013, p. 63). Depending on the success of carsharing services, in the end depending on whether there is a change in overall mobility patterns and the dominance of the privately owned car, the environmental benefits can be large, since it is estimated that one shared car substitutes for roughly 16 private cars (Canzler & Knie, 2011, p. 115 f.). 4.3 The Transformative Capacity of the Electric Vehicle Based on the ideal-type vision of e-mobility as a radically different sustainable form of future mobility, the question is how can this desired future be achieved? Throughout chapter 3 it has been shown that there is no straightforward way of managing transitions and that current developments and potential for system innovations need to be analyzed, in order to find ways of steering and modulating ongoing dynamics in desired directions. In a technological field such as e-mobility, processes of socio-technical co-evolution are at the heart of a wider transition. For a better understanding of socio-technical co-evolution it has been proposed to analytically separate the transformative capacity of new technologies and system adaptability. A new technology may have a high transformative capacity, exerting direct and disruptive pressure on its social environment, or a low transformative capacity, having only an indirect impact and leaving basic structures of a socio-technical system intact. Thus, a technology can either amount to a radical innovation inducing substantial change or an incremental innovation fulfilling subsidiary or additional functions. The transformative capacity may depend on whether an endogenous technology (i.e. a technology developed within an established field), or an exogenous technology (i.e. a technology developed outside its eventual field of application), is concerned. Transformative capacity is an inherently relational category, because it depends not only on the specific characteristics of a technology but also on the respective social structures on which it has an impact. So, new technologies can be shown to have a transformative capacity and thus a direct impact on socio-economic and institutional structures, while at the same time, the degree to which those social structures are adaptable determines the actual impact a new technology can have. Nonetheless, the central idea of introducing transformative capacity as an independent analytical category 165

179 (where, clearly, separation is possible only theoretically) is to be able to analyze the impact on socio-economic and institutional structures of a technology in its own right (Dolata, 2011, p. 12). In our case, the new technology is the battery-electric vehicle. This is not an unambiguous choice. One might argue that innovations in the field of ICT are much more relevant for triggering advancements in e-mobility and enabling integrated and intermodal forms of e-mobility. Thus, it is indeed possible to argue that this is where the actual transformative capacity is located. Similarly, in e-mobility a wide range of vehicle types are being developed and especially pedelecs and e-scooters are much more successful than electric cars (e.g.in touristic applications in Germany or as increasingly popular means of every-day transportation in Asian mega-cities). Furthermore, these vehicles play an important role for more sustainable forms of mobility and multimodal individual transport. Nonetheless, the BEV has been chosen as the relevant new technology even though it is not new in terms of its technological innovation and development, but has newly gained prominence for a number of reasons. The car, the automobile, is the central technological and especially also culturally relevant artifact when talking about personal mobility. Consequently, the emergence and the potential of e-mobility are publicly and politically debated in terms of the electric car, while all other aspects, e.g. other types of vehicles, transport modes etc., are variants of or elements related to the BEV. Thus, the electric car is the boundary object around which the field of e-mobility has evolved and which is central for all kinds of involved actors either as a technological solution, the starting point or an element in a new system, or an evil that needs to disappear. Therefore, the BEV is, after all, the central technology in relation to which different actor groups position themselves, develop attitudes, reactions and strategies, eventually determining overall system adaptability. So, how can the transformative capacity of the electric car which currently has only a marginal real impact be assessed? The approach chosen here is to identify the mis-fits of the electric car with the existing system structure, thus what is discussed as the technological shortcomings of the electric car in current debates. It is the mismatch between the potential of a new technology and the feasible possibilities of using it under given circumstances which can be characterized as the transformative capacity inherent in a new technology. The more a new technology becomes relevant for a sector and the less it can be embedded in existing sectoral structures, the larger is its transformative capacity and thus the pressure for change it exerts on established sectoral organizations, institutions and behavioral patterns (Dolata, 2011, p. 68 f.). The empirical consequence is in this sense a research approach that focuses on 166

180 the technological shortcomings that are attributed to a new technology. Technological shortcomings understood as mis-fits of a new technology with an established system are, in principle, shortcomings only with respect to the existing conditions and thus may in some cases very clearly point to the new technology s transformative capacity. In the case of e-mobility, such an approach may help to overcome a perspective where the mis-fit between a new technology and the existing system structures is discussed in terms of necessary but lacking technological improvements only. What is perceived as a technological shortcoming from the point of view of the current mobility system may have completely different implications for a potential future system that functions based on different infrastructures, use patterns, business models and is operated by different actors. Since it is hardly possible to identify unambiguous and universally valid indicators of transformative capacity, it cannot be measured quantitatively in any meaningful way. However, a number of qualitative criteria have been identified for assessing the characteristics of a new technology with regard to the specific configuration of the socio-technical system that is to be studied. As introduced in chapter 3.3.3, the transformative capacity of a new technology can be determined by analyzing in what ways it: alters the technological profile of a socio-technical system, affects established patterns within a socio-technical system (market relations, R&D, policy making, user behavior), facilitates new patterns of interaction between actors, initiates institutional readjustments, opens up or widens existing borders of the socio-technical system. With regard to the technological profile of the socio-technical system that has evolved around the fulfilling of mobility needs, a widespread diffusion of the BEV would entail major changes. Especially since Germany has a large automotive industry with a focus on the premium segment and powerful internal combustion engines, the BEV means a severe turning point and a restructuring of the automotive sector across its entire value chain (Hüttl et al., 2010; Kampker et al., 2013a). The modern automobile as the dominant means of transport has specific characteristics making up the basic features of its technological profile, such as being a multi-purpose vehicle, built from steel and powered by an internal combustion engine (Orsato & Wells, 167

181 2007, p. 995). OEMs and suppliers have developed competences accordingly and the introduction of BEVs would require completely different and new competences. Current activities in the fields of producing motors, gear units, exhaust systems and oil pumps become obsolete when the internal combustion engine is replaced by an electric motor and a battery. If OEMs want to retain their current value creation levels, large parts of battery production and production of components for electric motors need to be done by them directly, rather than sourced out or bought from suppliers (Gnann & Plötz, 2011, p. 38; Kasperk & Drauz, 2013b, p. 37 f.). Furthermore, these technologies are well-developed in other fields, but still have to be adapted to automotive applications. This is most obvious for the case of batteries. In general, electronic applications and nanotechnologies will play an increasingly important role, while mechanical and hydraulic components will disappear (Kampker et al., 2013b, p. 22f.; Kasperk & Drauz, 2013b, p. 37). These shifts in the technological profile also means that completely new resources are needed in the production process, especially rare earths, which have to be procured and processed differently (Rammler, 2011, p. 20). Such a fundamental shift which would be challenging for any industry is especially severe for the automotive industry, because OEMs have focused on increasing the variety of their offer by developing more vehicle types and by differentiated models within a specific series. This has led to increasing costs, shorter model and development cycles, which forces OEMs to concentrate on core competences and a limited amount of elements of the entire value creation process (Kasperk & Drauz, 2013b, p. 37). Such a concentration process limits flexibility and the ability to adapt to fundamental technological changes. Similarly, it is always difficult to fundamentally change elements that are closely related to the core of a specific brand. For the automotive industry as a whole, the core of a brand is the internal combustion engine, which is the distinguishing feature across the different car manufacturers and the central element in automotive value creation. Thus, the emergence of electric cars as the new core technological artifact would involve a fundamental re-structuration of the automotive industry in terms of re-defining core competences and the technological core of an automotive brand. Another fundamental uncertainty in such a transformation process relates to the question whether OEMs, suppliers or battery producers will develop the core competences for manufacturing electric drivetrains (Kasperk & Drauz, 2013a, p. 110 f.; Rammler, 2011, p. 20). A brief look into the history of the automotive industry and its origins at the turn of the 20 th century shows how path-dependencies have developed and led to a technological lock-in of the automobile powered by an internal combustion engine. Such a historical perspective 168

182 also helps to understand how severe a transformation towards e-mobility would be on various, not only technological, levels. In their historical analysis of the automotive industry, Orsato and Wells (2007) explain how the industry ended up being locked into a business model that gives primacy to high output, low unit market prices, and deriving revenues from the outright sale of new cars (including financing) rather than other income streams (p. 996). A number of factors and processes contributed to this outcome, for instance the emergence of Fordist production techniques and robust steel body-chassis technology, which together facilitated largely automated mass-production of cars. At the same time this increased capital intensity of automotive production and a focus on achieving economies of scale (Orsato & Wells, 2007, p. 995). This type of production optimally (and only) fits with a business model of selling large numbers of cars at generally affordable prices. The technological lock-in of the internal combustion engine was, in addition, partly resulting from efforts of oil companies, most prominently Rockefeller s Standard Oil. After the diffusion of electric lighting had led to losses for the oil industry, they lobbied car manufacturers to develop internal combustion engines in such a way that they would run on hydrocarbon fuels. The oil industry had a strategic interest in establishing direct links with the automotive industry as a new business opportunity and eventually succeeded in this endeavor (Orsato & Wells, 2007, p. 996 f.). All of these long-established principles and relationships would become obsolete, if the core of the automobile s technological profile disappeared in the process of electrification. This would also call for the development of new business models. Just like the ICE car, the electric car has its own technological specificities that will not fit with old principles of cheap mass production and realizing economies of scale (at least not in the beginning) and requires other strategic partners than the oil industry. Especially the high production costs and problems of range and charging time of electric vehicles which are the basic differences with regard to the technological profile of ICE and electric cars indicate that automotive value creation in an electrified age would look very different from today. Due to its range limitations, the electric car is no longer a multi-purpose vehicle, which implies that the industry would shift its focus from vehicle sales and financing to offering mobility services (Canzler & Knie, 2011, p. 77; Kasperk & Drauz, 2013b, p. 35). This is a fundamental and serious challenge, not only for the automotive industry, but also for customers: while it is rationally possible to question the reasons for consumers to keep buying over-dimensioned and over-specified cars (Orsato & Wells, 2007, p. 997), it is also obvious that the technological profile that has become dominant in the field of automobility is the norm against which the electric vehicle and concepts of e-mobility are being compared. 169

183 The obvious and substantial mis-match the electric car represents with regard to the established technological profile in the socio-technical system of automobility is reflected in the way that the automotive industry deals with the trend towards electrification. Great efforts are undertaken to develop electric cars that can compete with conventional cars in terms of price and range, and which are in essence and functionality not really different from the automobiles we have today (Wells & Nieuwenhuis, 2012, p. 1686). This strategy becomes obvious when looking at the principle of conversion-design which dominates the development of electric cars by established OEMs. The basic idea is to convert a conventional car into an electric car by simply exchanging the drivetrain (Kampker et al., 2013b, p. 22). This can shortly be summarized as an attempt to force this competing and quite different technology into the constraints of the existing regime to which it is not optimally adapted (Wells & Nieuwenhuis, 2012, p. 1686). Assuming that the automotive regime would be able to adapt to electric vehicles instead of the other way around the characteristics of the electric car could be the core of an overall fundamentally different technological profile of automobility. This would possibly lead to new business models as well. Some car manufacturers have already been experimenting with leasing concepts where batteries of electric cars are leased to customers, in order to lower the purchase price of the car (Wells & Nieuwenhuis, 2012, p. 1686). With regard to the electric vehicle itself, the approach of purpose design in contrast to conversion design aims at developing new vehicle concepts that utilize the specific advantages of electric vehicles, instead of trying to imitate a traditional car, which will most likely result at best in a less good alternative (Kampker et al., 2013b, p. 22). So, when the electric car s mis-fits with the current system would turn into the corner stones of a new system of automobility, this process could be described in terms of a system innovation. The fact that the electric car presents a mis-fit is alone an indication of its transformative capacity. The automotive industry is currently struggling with this situation. OEMs are hesitant to invest in the development of electric vehicles, because this is risky in a situation of uncertainty and where it is quite obvious that electric cars are not competitive on current markets. At the same time, political and economic pressure force OEMs to look for innovations and alternatives, with electric vehicles currently being presented as a promising option (Kasperk & Drauz, 2013a, p. 104 f.; Wells & Nieuwenhuis, 2012, p. 1688). The basic dilemma for the automotive industry is the fact that multi-purpose cars powered by internal combustion engines have been perceived to be the core of the automobile system s technological profile. ICE technology is consequently viewed as the core 170

184 competence of OEMs. The electric vehicle makes this core element obsolete and thus puts into question the actual role and function of those actors. The transformative capacity of the electric vehicle therefore does not only relate to technological change, but to changing perceptions of the role of OEMs from manufacturers of propulsion system to a broader role of providers of mobility (Orsato & Wells, 2007, p. 1004). Thus, the electric vehicle affects not only industry structures and the technological characteristics of the central product, the car, but also patterns in markets, R&D, policy making and user behavior (acatech, 2010). The envisaged diffusion of electric vehicles would not only produce technological mis-matches for the automotive industry, but all kinds of institutional and behavioral mis-matches in various dimensions affecting different groups of actors. A system of automobility has historically emerged where the car is the central technological artifact but where other elements, such as roads, gas stations and a fuel industry, user behavior, car-related services, regulation and policy etc., have together evolved into a working system. All of these elements and the patterns of interaction between them will be affected when the core product is attempted to be replaced by a, in many ways fundamentally different, new technological artifact (Geels, 2012, p. 473; Markard et al., 2012, p. 956). Users are affected, because electric vehicles cannot be used in the same way as conventional cars. They perform much worse as regards range, price and charging (Vallée & Schnettler, 2013, p. 30; Kampker et al., 2013b, p. 20 f.). The charging infrastructure also has to be built up (Kampker et al., 2013b, p. 19). The service infrastructure around car use, e.g. garages, dealerships etc., is also not in place for electric vehicles, thus presenting another restriction for users (Vallée et al., 2013, p. 100). As already indicated, new business models are needed to deal with the high prices of batteries. New use patterns, such as carsharing and forms of intermodal mobility, could help dealing with the high selling price and limited range of BEVs as well as the lacking charging infrastructure (Kampker et al., 2013a, p. 1 f.; Kasperk & Drauz, 2013b, p. 39 f.). R&D in the field of automobility would also have to re-focus, in order to accommodate new technological fields of research, such as chemical and electrical engineering. Typical marketing-oriented research focusing on user acceptance and diffusion patterns also will not be sufficient with regard to electric vehicles, as long as conventional cars and their characteristics are the standard frame of reference. Finally, the electric vehicle as a technology at the interface of the automotive and energy sector cuts across political domains, e.g. energy and transport policies. Furthermore, 171

185 the fact that a charging infrastructure for electric vehicles needs to be built up involves challenges for policies in the field of city planning and spatial management. With regard to institutional readjustments, the electric vehicle in many ways does not fit the existing mobility system in terms of regulations and standards. While conventional cars can be re-filled at every gas station, the plugs and charging infrastructure for electric vehicles differ according to operators and in different countries. The operation of a public charging infrastructure also needs to be regulated with regard to parking spaces, reservations, and billing systems (Vallée & Schnettler, 2013, p. 32). Where innovative mobility concepts are developed for fostering e-mobility e.g. urban intermodal mobility services integrating different modes of transport infrastructures and management systems are needed for organizing user-friendly access (Vallée et al., 2013, p. 63). The BEV also points towards new patterns of interaction, for instance, cooperation between OEMs and battery producers (Gnann & Plötz, 2011, p. 24 f.), or mobility providers and energy suppliers (Kasperk & Drauz, 2013a, p. 123 ff.). Especially the long-established networks of OEMs and automotive suppliers would have to be re-adjusted if the industrial core of producing ICE cars would shift to the production of electric cars, or even more fundamentally, if mobility services were to become central in combination with electric cars. Already today, it is normal for OEMs to build strategic alliances with other OEMs or large suppliers, in order to realize synergies and economies of scale, especially in the field of R&D. This pattern of interaction would have to be intensified and broadened, if it were to be adjusted to the electric vehicle. Many, also smaller, suppliers have competences in the fields of battery technology, electric motors and light-weight design. Thus, joint activities with these sometimes smaller and new partners would be advantageous, while a potential threat exists with regard to the delineation of boundaries in these forms of cooperation. Since they deal with the new technological core of electrified drive trains, they are of high strategic importance and require a high degree of integration (Kasperk & Drauz, 2013a, p. 123 f.). Similarly, large automotive suppliers can already be observed to adjust their regular patterns of interaction by cooperating with new partners that can provide them with new competences in the field of electronic applications and battery technology. The expected shift in the value creation process from traditional vehicle technology to the electric drivetrain is seen as a chance by larger suppliers to expand their position in the value chain, which may eventually lead to new patterns of interaction between OEMs and suppliers in general (Kasperk & Drauz, 2013a, p. 125; Kasperk & Drauz, 2013b, p. 38 ff.). Most likely, OEMs will try to position themselves as system integrators being involved in all levels of value creation. This can be 172

186 achieved by building a broad network and engaging in cooperation with key partners. Especially small and innovative firms in the field of e-mobility that are in need of financial resources will most likely be acquired by OEMs, in order to gain new resources and competences, and to keep potential competitors in check (Kasperk & Drauz, 2013a, p. 123 f.; p. 126). Last but not least, completely new types of interaction will have to emerge, in order to deal with the most severe mis-fit of the electric car. It is no longer powered by fossil fuels, but by electricity, thus requiring new suppliers and a completely new charging infrastructure. Furthermore, replacing the ICE by an electric motor and battery leads to a much more limited vehicle range and relatively long charging times. This means that new patterns of interaction with energy suppliers would have to be found, in order to guarantee a sufficient build-up and operation of charging infrastructure. For energy suppliers entering the mobility market it is a challenge to find new patterns of interaction with their customers, because charging infrastructure operated by different energy companies and used by different types of private vehicle owners or service operators, e.g. carsharing firms, requires new offers and services, going beyond electricity being a relatively simple commodity supplied in one standard way. The integration of the vehicle and the charging infrastructure, the organization of charging and billing processes as well as the management of mobility services requires new patterns of interaction with the ICT sector as the third key industry linked with the automotive and energy sectors, in a new system that may evolve around the electric car (Kasperk & Drauz, 2013a, p. 126; Kasperk & Drauz, 2013b, p. 40). These new forms of interaction show that the borders of the socio-technical system will be opened in the course of developing e-mobility, for instance, by sectors entering the mobility field that have not been part of it before, e.g. energy, battery producers and ICT companies. Electrifying the drivetrain will not only lead to changes in the interactions between OEMs and suppliers or changes in components of automotive production. Rather, this would mean the emergence of a fundamentally different industry structure, where new actors become central and relationships are being redefined. Obviously, battery producers would play a central role and could position themselves along the automotive value chain (Kasperk & Drauz, 2013b, p. 41). More fundamental change is also likely. For instance, e- mobility is a completely new field of business for energy suppliers, which does not only offer an additional opportunity to sell electricity, but could also induce the development of completely new business models, which may in the long term facilitate a central role for energy companies emerging as providers of mobility solutions (Kasperk & Drauz, 2013a, p. 173

187 110). As already mentioned, ICT companies would come to play a central role as providers of the connecting link between the electric vehicle and the energy infrastructure in which it needs to be embedded. These companies may become a central new actor in an e-mobility system, because apart from providing a key technology, they might even emerge as the providers of intermodal mobility services (e.g. integrated service offer including reservation, charging, billing as a one-stop-shop solution) and the primary contact point for customers via Smartphone applications (Kasperk & Drauz, 2013a, p. 117 f.). In addition, the established actors in the field, the OEMs, could redefine their role, looking into stationary applications of batteries or energy services (as for instance done by Volkswagen AG in cooperation with the energy supplier Lichtblick SE), or investing themselves in energy production and charging infrastructure. They might also cooperate with energy and ICT companies as providers of mobility services, utilizing their established after-sales and services infrastructure in innovative ways (Kasperk & Drauz, 2013b, p. 41 f.). In general, the electric vehicle as the central technological artifact of a potential new e- mobility system questions the role and functions of established actors and of the potential newcomers. Depending on how the different actors position themselves in this new field, how they manage to demarcate their individual claims, and find ways of cooperating with each other, a fundamentally different socio-technical system might emerge around the electric vehicle (Kasperk & Drauz, 2013a, p. 115 f.; Kasperk & Drauz, 2013b, p. 45). The fact that a large number of non-automotive actors is from the beginning involved in or someway related to the development of e-mobility may open up new perspectives on (electric) cars, the way they are produced, designed and used (Vallèe et al., 2013, p. 90). In this way, the electric vehicle possesses an indirect transformative capacity: by triggering shifting system boundaries and new forms of cooperation, the conditions for its own further development and the way it becomes embedded in its social environment are shaped by actors with a fresh perspective on mobility. New perspectives and the combination of different perspectives in a new field are potentially important for the development of e-mobility as a system innovation, because change in basic interpretive schemes, guiding principles and norms, requires processes of reflection and learning. Tbl. 4 summarizes the findings regarding the transformative capacity of the batteryelectric vehicle. This brief overview indicates that the transformative capacity of the BEV is high. While these criteria of transformative capacity need to be studied in more detail, in order to understand the concrete implications for industry, policy, markets, regulations etc., the point here is to show the very basic mis-fits of the BEV with the current mobility system. 174

188 BEVs cannot simply replace conventional cars without major changes in industrial production structures, business models, use patterns and political regulations. Whether these mis-fits with the current mobility system unfold a potential for more systemic types of change, depends on the degree to which the socio-technical system around it can be characterized as adaptable. Tbl. 4: The transformative capacity of the battery-electric vehicle. Some general observations can be made that indicate that the established sociotechnical system revolving around automobility shows signs of growing system adaptability. In theory, the key elements determining whether overall system adaptability is low or high are reflected in regime configurations that are either well-aligned or characterized by internal tensions. Geels (2012) has identified four cracks in the automobility regime that are indicators of destabilized regime structures and thus potentially factors increasing overall system adaptability. These are, first, increasing traffic-related problems in cities, such as congestion and local pollution, which decrease quality of life in cities at a time when increasing numbers of people live in urban areas. Second, the market for cars in most developed countries, especially in North America and Europe, are saturated as can be shown by decreasing volumes of car sales, indicating that peak car has been reached (Goodwin, 2012, p. 11; Newman & Kenworthy, 2011). This aspect is also discussed by Orsato and Wells (2007), arguing that there are many reasons to believe that the foundations for this historic success [of the automotive industry] are fading away (p. 998). It is shown that the automotive industry is in an overall difficult situation with saturated markets in developed 175

189 countries, unexpected difficulties in entering emerging markets (e.g. due to price instabilities and lacking infrastructure), and a business model based on the need to realize economies of scale to deal with low profit margins, while customers demand a wide variety in products. In combination, these factors have at least increased the overall vulnerability of the historically successful automotive industry (Orsato & Wells, 2007, p. 999). Third, political pressure on the automotive industry and on unrestricted car use is slightly increasing. Some efforts are undertaken with regard to the regulation of CO 2 -emissions, local traffic management and a broader orientation along the principle of sustainable mobility. This crack is to be treated cautiously because overall the automotive industry is not severely threatened by political regulation as has been shown by governmental support during the financial crisis or successful lobby efforts by the German automotive industry in EU negotiations on emission regulation. Fourth, there is an increasing awareness of environmental problems and resource scarcity. This crack also needs to be treated cautiously, because especially for OEMs this is not so much a direct problem, but rather an indirect one, articulated by politicians and to a lesser degree by customers (Geels, 2012, p. 478 f.). Departing from these cracks as macro-level societal and economic trends, some early signs of minor change in the automobility world can be observed. An example is the (eventually failed) business model of Better Place, a mobility company, which focused on operating charging spots and battery exchange stations and leasing batteries to customers, thus reducing their initial purchase cost of a BEV (Christensen et al., 2012, p. 501 f.; Orsato et al., 2012, p. 214 f.). Thus, in this case some of the most prominent problems with BEVs, namely battery costs and recharging infrastructure, have been addressed in the form of an innovative business model. A second example is the problem of limited range of BEVs and planning challenges when BEVs are to be used for electricity storage. These technical problems can be dealt with much better by fleet operators or carsharing service providers than by individual car owners. Thus, carsharing companies and fleet operators may not only have an impact on the development of e-mobility as users but potentially also introduce new use patterns of mobility that are not centered around private car ownership (Orsato et al., 2012, p. 217 f.). Thus, referring to the basic category of transformative capacity of new technologies, it can be shown that the alleged technological shortcomings of battery technology are in some cases a trigger for change in social processes and thus open perspectives for a potential system innovation around e-mobility in the future. However, despite a few innovative approaches and even though cracks can be shown to have appeared, the overall automobility regime still seems to be relatively stable. The focus 176

190 of OEMs is not so much on experimenting with innovative business models than it is on trying to develop electric vehicles that can be compared to and compete with conventional cars. The basic approach of industry R&D is geared toward a technological fix of environmental problems possibly arriving at an economic and environmental win-win outcome. Dealing with the more fundamental, underlying challenges of a car-centered system is thus effectively being avoided (Nykvist & Whitmarsh, 2008, p. 1377; Wells & Nieuwenhuis, 2012, p. 1687). Similarly, studies of user acceptance of electric cars show that in general people are interested in the new technology and in buying electric vehicles as long as they function basically like a normal car. The conventional car is the commonly known standard and electric vehicles are evaluated by directly comparing them in the light of what is known. Then, obviously, the electric vehicle does not perform sufficiently with regard to range, price and charging time. Interestingly, the few aspects where the electric car performs better, e.g. offering the possibility for charging at home, a new type of driving experience and being locally emission-free, are seen as much less relevant (Schlager & Oltersdorf, 2011, p. 133). Comparing electric vehicles with conventional cars as the standard, especially in terms of range and price, is also the general pattern observed in political efforts and in the field of (publicly funded) R&D, where the majority of financial resources are dedicated to battery technology. The battery is perceived to be the basic technological weakness, because it is still too expensive and not performing well enough to allow for an acceptable vehicle range (Weider et al., 2011b, p. 105). These examples show that system adaptability must be more than the existence of cracks in a regime, which then serve as a window of opportunity for new technologies to emerge and diffuse, resulting almost automatically in an ensuing process of socio-technical co-evolution. System adaptability is an inherently agency-related category. Regime structures are more or less actively and consciously reproduced and drawn upon by actors in their realworld contexts. Thus, the emerging cracks as an analytical category may give hints regarding instances for potential system adaptability at an aggregated level. However, whether or not a concrete system is adaptable depends critically on the way that actors perceive and deal with change, and not just the fact that on a macro-level things change, trends and crises occur. Similarly, the transformative capacity merely points to the inherent potential of a technology in relation to given regime structures, thus as a concept it captures the virtual range of possibilities or alternatives that become feasible with this technology. 177

191 Actual change results from the interplay that unfolds within the social configuration of a system and its constituting actors. For the case of e-mobility this means that its potential can only show where it is applied in practice and where experimentation and learning can take place. Only when there is actual, micro-level socio-technical interaction, is it possible that the electric vehicle gains meaning and a symbolic value of its own, independent from the conventional car as standard frame of reference. This will then also prepare the ground for new mobility patterns, industry structures and business models (Weider et al., 2011b, p. 106 f.). The cracks and macro-level perspective described briefly in this section do show that the automobility regime is facing severe problems and that there is a potential, or a threat for many, of structural change. At the same time, there is no convincing and completely spelled-out alternative waiting to replace elements of the old system. Most likely, a transitional phase is beginning to take shape, where it is still a major challenge to begin to re-think mobility (Canzler & Knie, 2011, p. 31). Therefore, in the following, a more detailed look is taken at the German innovation system, where experimentation in the field of e-mobility is actually taking place. A focus on this lower level (as compared to macro trends in the automotive industry and society as a whole) allows for understanding micro-level socio-technical co-evolution and behavior of actors involved in developing e-mobility. The aim of this endeavor is to better understand where the actual seeds for system adaptability may be found, or what the major barriers are, which hamper the development of e-mobility even in a situation where framework conditions are favorable. 4.4 Case Study: E-mobility in Germany Assessing System Adaptability Since the new technology, in this case the electric car, can be assessed regarding its basic technological properties and its envisaged function in the context of the ideal-type system innovation, at this point the potential of such a system innovation actually evolving depends on system adaptability. In the following, a case study on the German innovation system for e- mobility will be presented that aims at shedding some light on the concept of system adaptability and what it entails in this specific case 6. Preliminary results of this case study have been published in Augenstein (2014, forthcoming). 178

192 4.4.1 Case Study Design The aim of the case study is to analyze system adaptability in the field of e-mobility development specifically in the German state of Baden-Württemberg. Since system adaptability is characterized by changes in rules and resources and depends on actor and power constellations, the case study has to include a combination of methods that can grasp these different aspects. Before these concrete steps of data collection and analysis are layed out, the boundaries of the case have to be defined. With regard to the period of investigation, the year 2009 has been selected as the starting point, since this was the start of the first major funding program Model Regions for Electric Mobility, and the subsequent developments are followed up until the present. With regard to the spatial boundaries of the case, an embedded design has been chosen. The central unit of analysis is the German state of Baden-Württemberg, which as a federal state, i.e. one of the German Länder, has clear geographical and political boundaries. However, since the funding programs for e-mobility are launched by the federal government and the overall strategy for developing e-mobility is also shaped at the federal level, the case of Baden- Württemberg needs to be embedded in its broader German context. Therefore, chapter will briefly describe the German innovation system for e-mobility as a whole, with a particular focus on political initiatives at the federal level, the development of funding programs and the actor structure determining the strategic perspective of the German government. Relevant data sources are official government documents and existing studies on the goals and actor structure of the German innovation system. These are screened, in order to identify power structures and the dominant rule systems influential actors refer to. In order to gain an in-depth insight on system adaptability, the analysis also needs to focus on concrete interactions of different groups of actors, which can best be observed in the context of concrete e-mobility projects carried out in Baden-Württemberg (chapter 4.4.3). Against the background of the overall actor structure, dominant rule systems and patterns of resource allocation in Baden-Württemberg as a whole, these individual projects can be understood as sub-cases embedded in the broader context of Baden-Württemberg s innovation system (which is in turn embedded in a broader German context). Collecting and analyzing data within these sub-units of analysis can help to illuminate the case as a whole (Baxter & Jack, 2008, p. 550), thus insights on system adaptability for e-mobility in Germany. This case study design entails thus a typical case-oriented, rather than variable-oriented, analysis where insights within the different sub-cases will be compared and related in the end, in order to 179

193 identify general patterns or dynamics (Kohn, 1997, p. 5). The results of analyses at these different levels will be re-integrated in chapter Following the broad overview of the German innovation system in chapter 4.4.2, which will mainly be based on official government documents and secondary sources, e.g. the few existing studies on the German innovation system for e-mobility, the analysis for Baden- Württemberg will proceed as follows. First, a brief overview is given on the region of Baden- Württemberg in general and the e-mobility initiatives taking place during the period of investigation from 2009 until today (chapter ). Second, a network analysis will give an overview of the actor structure of Baden- Württemberg s innovation system (chapter ). The method of network analysis is suitable for analyzing system adaptability, i.e. the degree to which a system is well-aligned or showing regime cracks with regard to dominant rule systems and actor-/powerconstellations: Basic premises of social network analysis are that actors in a specific network are interdependent and their relationships can be captured in terms of structural properties, which can be both constraining and enabling (Jansen, 2006, p. 18 f.; Wassermann & Faust, 1994, p.4). Thus, the basic conceptualizations on which network analysis builds is well in line with the concept of system adaptability in terms of regime structures. Social network analysis has a rich history, it is applied across different disciplines, it combines qualitative and quantitative approaches in widely varying degrees of sophistication (cf. Scott, 1991; Wassermann & Faust, 1994). For the purpose of this case study, where the relevant network is relatively small and the aim is to identify actor constellations in relation to their resourcefulness or influence on a regional innovation system and the basic rule systems (frames of reference and interpretive schemes in the context of e-mobility), a relatively simple network analysis will suffice. It will proceed as follows. As a starting point, relevant actors making up the innovation system are identified based on their participation in projects and receiving R&D funding in the context of the official governmental programs, or other forms of participation in institutionalized settings of the innovation system, e.g. specific sectoral or regional networks or associations. Relevant data sources are thus official documents, e.g. an overview of recipients of governmental funding published by the Federal Ministry for Education and Research (BMBF), or information brochures published by the Regional Project Coordination Agency (Regionale Projektleitstelle, PLS) in Baden-Württemberg. The identified actors are then clustered according to their respective affiliation with a specific industry or sector, the field of science and research, the public sphere or civil society organizations. An overview is thus gained on the basic actor structure of the network. In order 180

194 to analyze dominant rule systems, including an assessment of who shapes them to significant degrees, and patterns of resource allocation, i.e. overall power relations, in a next step, the relations and power positions of actors within the network are studied. The actors are grouped according to the relative influence they have on the innovation system as a whole, which can be low, moderate or high. Indicators for influence are the amount of funding an actor receives, the number of projects in which an actor participates and whether or not an actor functions as project leader in larger project consortia. It is assumed that these criteria are appropriate, because receiving financial resources from the federal government signals an important role attributed to a specific actor group and their specific focus of activities, and thus dominant rule systems with regard to perceived key aspects of developing e-mobility. Being involved in one or more concrete projects, and especially leading a project, signals the availability of both authoritative and allocative resources, and it grants the respective actor the possibility to actively shape strategies of developing e-mobility. The broad categorization of levels of influence is thus determined by the relative position of actors compared to each other. Finally, the network analysis focus shifts beyond the level of individual actors and includes a focus on the level of projects. Based on the relevant criteria for a sustainable system innovation developed in chapter 4.2, e-mobility projects carried out in Baden- Württemberg are clustered according to their specific focus, ranging from vehicle technology to system-innovative approaches. This overview provides insights on dominant rule systems and patterns of resource allocation, because it shows whether established power structures are reproduced in the form of conventional project designs, or whether substantial material and immaterial resources go to system-innovative projects. Finally, it will be analyzed what types of actors, with regard to affiliation and levels of influence, typically participate in what types of projects, in order to identify typical patterns of cooperation and respective outcomes relevant for the potential of a system innovation. Third, discourses that reflect the basic frames of reference dominating the network are analyzed, in order to complement the insights on different levels of influence of specific actors with their respective views, guiding principles and resulting strategies (chapter ). An overall assessment of system adaptability is then possible by identifying the dominant rule systems of different actor groups and the relative influence that these have, based on the power position of their respective proponents. Relevant data sources are official statements or position papers. A focus of the analysis will be on tensions or contradicting lines of argument and the relative position of varying frames of reference with regard to the influence of the actors that can be related to different positions. 181

195 Fourth, a closer look will be taken at projects that can be characterized as systeminnovative (chapter ). These projects are selected based on the criteria identified with regard to an ideal-typical system innovation: project focus on intermodality, carsharing, integration with renewable energy, and, in general, a focus on mobility that is not carcentered, as e.g. in a project on e-mobility and housing. These projects are analyzed with regard to their specific focus and design. Based on the overview of the actor structure across all the projects in the previous section, a more in-depth analysis of the actors and cooperation patterns involved in these projects is carried out. In-depth interviews 7 have been carried out with actors responsible for managing these projects. These were the project managers of involved firms (in the fields of energy supply, public transport, ICT, housing) and researchers. Interviews have also been carried out with representatives of the Regional Project Coordination Agency (PLS), i.e. the public agency responsible for coordinating the different projects. The semi-structured interviews, ten in total with each lasting between 60 and 90 minutes, focus on general perceptions and visions of e-mobility as well as the concrete projects and the funding program in general, actor constellations and conflicts 8. Thus, the role of the interviewed actors is twofold: On the one hand, as project partners involved in systeminnovative projects they are part of the innovation system as the object of study; on the other hand, they are also experts with an insider perspective on the innovation system as a whole and treated simultaneously as sources of information about the object of study. This is accounted for by separating these two roles into different sets of question, either directly related to the specific projects and personal perceptions, or focusing on assessments of developments in the field of e-mobility in Baden-Württemberg and Germany. Obviously, there may still be issues of overlapping roles, for instance, assessments of the innovation system structure are influenced by personal experiences in a concrete project setting, or personal beliefs regarding the overall development of e-mobility have an impact on how the innovation system is assessed. This ambiguity is dealt with by carefully analyzing the interviews based on theory-informed categories of perception of and ways of dealing with e- mobility (system adaptability) along the criteria of transformative capacity. 7 To ensure openness during the interviews, anonymity has been guaranteed to the respondents and their names as well as those of their respective organizations are withheld in this thesis. 8 The guiding questions used in the interviews have been developed based on the theoretical concepts developed in ch and the detailed questionnaire can be found in Appendix D. 182

196 Finally, the results of the case study of Baden-Württemberg are discussed and results are being presented in chapter , while chapter discusses the findings in the context of the German innovation system as a whole. Fig. 11: Case study design The German Innovation System for E-mobility As in many other countries, there has been more than one attempt to introduce electric vehicles and develop e-mobility in Germany. Some of the basic technological innovations, i.e. electrical motors and various types of batteries, date back to the early 20 th century and in more recent history, the 1990s have been a major hype phase (Høyer, 2008, p. 70 f.). Before this more substantial hype, the increasing awareness for global environmental problems, local pollution caused by transport especially in large cities as well as the oil crisis during the 1960s and 1970s spurred a phase of intense research and development of electric vehicles. When the dependence on oil from OPEC countries emerged as a real threat and stricter environmental regulations were adopted and against the background of decreasing growth rates in car production two large German OEMs, Daimler-Benz AG and Volkswagen AG, bundled their efforts in developing electric cars. However, these efforts did not materialize and visions of electric mobility vanished again (Canzler & Knie, 2011, p. 102 f.; Høyer, 2008, p. 66). More substantive action was taken during the 1990s. The general sustainability discourse had gained momentum and an important factor was the Zero Emission Vehicle regulation adopted by the California Air Resources Board (CARB) (Dijk et al., 2013, p. 136; 183

197 Sierzchula et al., 2012, p. 50). In Germany, a large-scale demonstration project was launched in 1992 by the German Federal Ministry of Education and Research (BMBF) and the Federal Ministry for the Environment, Nature Conservation and Nuclear Safety (BMU). A fleet trial with 60 electric cars was carried out on the island of Rügen. Apart from that, a number of smaller pilot projects aimed at market introduction of electric vehicles in some of the German states ( Länder ). The results of these projects which involved only the automotive sector as industrial partners were discussed controversially, especially political actors voiced ecological concerns regarding significantly higher CO 2 -emissions of electric vehicles as compared to conventional cars, due to the German electricity mix at that time. Also, in California the political regulation on zero emission vehicles was subsequently weakened and in the end basically nullified as a result of legal action initiated by different US car manufacturers. So, by the end of the 1990s, German activities in the field of electric mobility had come to a relatively abrupt halt (Canzler & Knie, 2011, p. 103; Schwedes, 2011, p. 12 f.; Weber & Hoogma, 1998, p. 558; Weider et al., 2011b, p. 108). The most recent hype around electric mobility took off in A number of framework conditions have been relevant for this renewed interest: Climate change had resurfaced as an important issue on political agendas and gained prominence in public as well as scientific discourses. Especially in the fields of energy and transport the heavy dependence on oil and the threat of Peak Oil have been increasingly perceived as a serious problem (Schwedes et al., 2013, p. 75). Furthermore, the global financial and economic crisis had negative impacts on the automotive industry, which already has to deal with saturated markets in most industrialized countries and is faced with a constant pressure to reduce costs due to its current system of production (Orsato and Wells, 2007, p. 998) Political Initiatives by the Federal Government: Caught between Environmental and Economic Goals Against this background, the German government stepped in to support the nationally important automotive industry (Schwedes et al., 2013, p. 75). The government s focus on e- mobility as the favored technological option can be explained by a number of aspects. With the German energy transition underway, BEVs were seen as important means of energy storage in a renewable energy system. Furthermore, advances in battery technology, concerns regarding the future of automobility in global mega-cities and especially the growing importance of the Chinese market for German OEMs and China s efforts of fostering the electrification of the car played an important role, too (Dijk et al., 2013, p. 141 ff.; Schwedes 184

198 et al., 2013, p. 75 f.). The German government also has to deal with political pressure at the EU level to achieve the 20/20/20 goals of increasing energy efficiency and the share of renewable energy by 20% until 2020, to contribute to the general vision of sustainable transport development (EC 2001; EC 2011) and the focus on developing alternative vehicle technologies in the 7th Framework Program for Research (Dijk et al., 2013, p. 138; Köhler et al., 2009, p f.). The attitude of political actors had thus changed compared to the 1990s. In 2007, the German federal government adopted an Integrated Energy and Climate Program (IEKP), which contains measures and draft legislation for increasing energy efficiency and fostering the development of renewable energies. The development of electric mobility is included as one measure for reducing CO 2 -emissions in the transport sector (BMWi & BMU, 2007, p. 89). For the first time, electric mobility has been taken up by political actors as an element of strategies for achieving a transition towards renewable energies and a contribution to climate change mitigation (Schwedes, 2011, p. 14). Electric mobility was also increasingly seen as an economically important field for technological innovation. The German government launched the Economic Stimulus Package II as a way to react to the economic and financial crisis in 2008/2009 and it contained, amongst others, measures for fostering research and development of electric mobility worth 500 million (Schwedes, 2011, p. 14). This combination of environmental and economic goals pursued by the German government in the field of electric mobility remains a central characteristic of the strategic initiatives taken in the following years. In 2009, the German government launched its National Development Plan for Electric Mobility (NEPE), which apart from setting goals of climate production and reducing dependency on oil, spells out the aim for Germany to become a lead market for electric mobility and to have one million electric cars on Germany s roads by 2020 (Bundesregierung, 2009, p. 46). To reach this goal, a comprehensive strategy ranging from basic research to market introduction is to be developed and actors from science, industry and politics are to be included in the strategy development process. Important elements would include not only the entire value chain around the electric car, but also its integration in the energy system and especially its potential contributions in the context of the transition to renewable energies (p. 4). The necessary R&D efforts in the fields of battery development, components and standardization, grid integration, battery recycling, information and communication technologies, education and competence building as well as preparing the technology and the market for commercialization are to be financed by the 500 million of the Economic Stimulus Package II implemented to deal with the 185

199 economic crisis in 2008/2009 (p. 24). In the context of the Economic Stimulus Package II, the Federal Ministry of Transport, Building and Urban Development has launched Electric Mobility in Model Regions, where eight regions have been selected as model regions for the period from 2009 to 2011 (Tenkhoff et al., 2011, p. 11 f.). The goals envisioned in the national development plan are specified further in the Government Program Electric Mobility launched in 2011 (Bundesregierung, 2011, p. 5). The government program is meant as the starting point for a new phase of launching a market for electric mobility, introducing concrete measures following the preparatory phase starting in This is seen as an important milestone on the way towards Germany becoming a lead market for electric mobility. The necessary measures are consequently embedded in a market logic of supply and demand: R&D efforts should concentrate on drive technologies and components as well as the integration of vehicles in energy and transport systems while it is emphasized that these efforts should be bundled in overarching thematic research projects. The basic rationale for this is to support relevant German industries and to be able to showcase the quality of German products in newly emerging systems for electric mobility. The demand side is addressed in a similar way. Research projects on the integration of electric vehicles in energy and transport systems are basically needed as a means of creating visibility and of demonstrating e-mobility to users (Bundesregierung, 2011, p. 7). However, it is also emphasized that establishing Germany as a lead market for electric mobility implies more than developing new drive technologies, but rather aims at future more sustainable energy and transport systems. Special emphasis is therefore put on the changing relations between established industry sectors, changes in individual lifestyles and the role of cities and urban infrastructures. New forms of cooperation will be needed on the way towards the lead market; however, the cooperation viewed as central in the government program are industry coalitions along new value chains, which should create the market, while science and policy function as innovation support and framework provision (Bundesregierung, 2011, p. 15). A number of concrete measures are envisaged in the government program. First, a joint R&D program is to be launched to speed up market breakthrough of electric mobility. There have already been the 500 million from the Economic Stimulus Package II that have been invested during the period from 2009 to Following this, another billion of governmental funding is made available until It will be used to finance large demonstration projects, the so-called Technological Lighthouses for Electric Mobility and Electric Mobility Showcases. R&D efforts should concentrate on the identified key 186

200 challenges with regard to battery technology, electric vehicle technology (focusing on materials, components and vehicle energy management, in order to realize cost effects), the charging infrastructure and grid integration. Furthermore, a coordination office, the Joint Agency for Electric Mobility of the Federal Government (GGEMO), is being set up to provide funding advice and coordinate the various R&D activities (Bundesregierung, 2011, p. 18 ff.). Second, the large demonstration projects envisaged in the funding program are highlighted (Bundesregierung, 2011, p. 25 ff.). They play an important role as new and innovative R&D instruments that are introduced in terms of a systemic approach. The socalled Electric Mobility Showcases are large-scale regional demonstration projects where the different industrial sectors and the public sector, in this case the Länder, work together to make electric mobility accessible to the general population and to make visible the technological competence of the German industry in a limited number of major projects. Consequently, the characteristics (and thus also selection criteria) for showcase regions are a systemic approach (especially including the interfaces between the electric vehicle, the energy and the transport system), cooperation across value chains, testing regulatory frameworks, reaching a critical mass to assess everyday suitability of e-mobility solutions, involvement of the public, a focus on education and professional qualification, and (financial) commitment of involved industry and public partners (Bundesregierung, 2011, p. 26 f.). In addition to the showcase regions there will be Technological Lighthouse projects, which will focus on distinct technologies and specific fields of application. They are seen as a strategically important complement to the showcases by developing technologies that can quickly be used in practice and contribute to achieving cost effects (Bundesregierung, 2011, p. 29 f.). Both programs are envisaged to mutually profit from each others experiences and results. Since only a limited number of showcase regions have been envisaged, it was hoped that some of the already existing model regions for electric mobility can be continued as showcases. The remaining model regions should, where possible, be continued in the context of the normal funding programs of individual ministries while the joint funding program focuses on the showcase regions (Bundesregierung, 2011, p. 28). Third, roadmaps and various governmental initiatives, e.g. organizing stakeholder conferences or engaging in EU and international political arenas, are envisaged for the fields of education and professional qualification for electric mobility (Bundesregierung, 2011, p. 31), questions of certification and standardization of products and production processes 187

201 (Bundesregierung, 2011, p. 32 f.), as well as international cooperation (Bundesregierung, 2011, p. 56 f.). Fourth, in the fields of charging infrastructure and energy provision (Bundesregierung, 2011, p. 34 ff.), and resources, materials and recycling (Bundesregierung, 2011, p. 41 ff.) as well as a number of legislative initiatives are planned and key challenges and open regulatory questions are identified. Furthermore, legal measures relating to road traffic law or emission and environmental law as well as (tax) incentives are laid out (Bundesregierung, 2011, p. 46 ff.). In sum, it can be shown that there is, overall, strong political commitment to developing e-mobility in Germany. A broad range of measures and substantial amounts of financial resources are dedicated to the goal of developing a lead market. However, it can be shown that under the surface, tensions are emerging between environmental and economic goals of developing e-mobility, and between classic industry R&D and more systeminnovative approaches, for instance in the showcase region program Actor Structure in the German Innovation System for E-mobility: A Key Role for Regime Incumbents Apart from setting strategic goals in the National Development Plan and providing governmental funding for R&D, the German innovation system for e-mobility has also been characterized by an institutionalization of organizational structures at the federal level. Early in 2010, the Federal Ministry of Economics and Technology (BMWi) and the former Federal Ministry of Transport, Building and Urban Development (BMVBS, re-structured in 2014 as Federal Ministry of Transport and Digital Infrastructures (BMVI)) set up the Joint Agency for Electric Mobility of the Federal Government (GGEMO), located at the BMWi. The GGEMO functions as a Secretariat which coordinates the activities of the federal government in the field of e-mobility, organizes official events and conferences, and provides support to another important body, the National Platform Electric Mobility (NPE). Based on the National Development Plan, the German government has invited representatives of industry, science, politics, labor unions and civil society to build this National Platform Electric Mobility in The members of the NPE have decided to follow an approach that is systemic, marketoriented and technology-neutral (NPE, 2010, p. 5). They have committed themselves to the 2020-goal of establishing a lead market and having one million electric vehicles on the road by Within the NPE, activities are focused across seven working groups dealing with the 188

202 following aspects: 1) power train technology; 2) battery technology; 3) charging infrastructure and grid integration; 4) standardization and certification; 5) materials and recycling; 6) education and qualification; 7) framework conditions. The NPE as well as the individual working groups regularly report on the progress that has been made and provide the government with recommendations (NPE, 2010). An overarching steering group meets regularly to discuss the progress that has been made in the individual working groups. Members of the steering group are the state secretaries of the involved federal ministries, the Joint Agency for Electric Mobility of the Federal Government (GGEMO) and the chairpersons of the seven working groups (usually members of company boards). Steering group meetings are also attended by the Industriekreis, a group consisting of representatives of industry associations (VDA, BDI) and labor unions (IG Metall). The official reports of the NPE are written by an editorial team, which consists of representatives from each of the working groups and the Industriekreis (Canzler & Knie, 2011, p. 88). The main topical focus and activities of the seven working groups show that the NPE is guided by a traditional approach where the problem is separated along disciplinary lines and is basically treated as a technological problem. It can be argued that this is the logical consequence of the member structure of the NPE. Representatives of the automotive and the energy sector have been dominating the working groups as well as formal and informal decision-making processes. Public transport actors were not represented at all and interests of civil society organizations and users in general remained underrepresented (Canzler & Knie, 2011, p. 11). A study carried out by the Independent Institute for Environmental Issues and the Institute of Transportation Design, which has been conducted on behalf of the Federal Ministry of Transport, Building and Urban Development (BMVBS, now BMVI), finds that the composition of the NPE reflects the general actor structure of the innovation system as a whole (Zimmer & Rammler, 2011). The aim of this study was to provide an overview of the relevant actors in the German innovation system and assess their impact on future technological developments as well as their guiding visions and perceptions of e-mobility (Zimmer & Rammler, 2011, p. 186). The study shows that two-thirds of the actors making up the innovation system are business and industry representatives. The last third is made up of scientists and politicians, while consumers and civil society organizations hardly play a role. It can be shown that only one third of all those actors have a significant impact on the structure and developments within the innovation system. These are representatives of OEMs and their supplier networks, large energy suppliers and some actors from the chemical 189

203 industry, electrical engineering and ICT and their respective industry associations. In the realm of politics, influential actors are the federal government and the involved ministries, and, in the field of science, a small number of universities and research institutes also have a significant impact. This actor constellation is in line with the innovation system s emphasis on building up the charging infrastructure and developing battery technologies (Zimmer & Rammler, 2011, p. 83 f.). The network analysis carried out in this study also shows that by launching the NPE and giving it a central role, the German government has actively granted the established industrial actors the opportunity to dominate developments. The leading positions in the NPE are held by representatives of the automotive sector, battery technology and engineering as well as large energy suppliers thus, actors subject to fundamental path dependencies and with a focus on product innovations. Consequently, the study also finds that those actors that could be expected to contribute to the development of system innovations (e.g. small firms in the fields of electric cars, charging infrastructure or renewable energies, actors at municipal levels) are lacking influence and the financial means to have an overall impact on the innovation system (Zimmer & Rammler, 2011, p. 84). Even though the NPE is made up of established industry actors, it became nonetheless obvious that developing e-mobility would amount to a challenge that goes beyond sectorspecific technological questions. Car manufacturers are aware of the fact that the global trend of growing mega-cities threatens their business model of selling cars and that they will have to find forms of cooperation with the public transport sector. Similarly, energy utilities have realized that e-mobility is not only about changes in the field of transport but that integrating electric vehicles and smart grids eventually implies a fundamentally different system of energy provision. Discussions in the NPE have shown that most of the involved actors have trouble coming to terms with redefining their established strategies, business models and patterns of cooperation, due to great uncertainties regarding future products, markets and power relations between the actors involved in developing e-mobility (Canzler & Knie, 2011, p. 11 f., 14). Conflicts also loomed with regard to the allocation of governmental funding. Energy utilities claimed the need to have the charging infrastructure subsidized, car manufacturers were arguing for subsidizing electric vehicles in order to decrease their purchase price, and the science partners also voiced their demands. It became clear rather quickly that the sum of these individual claims would be far beyond any realistic government program and that compromises would have to be found. A pattern emerged in these negotiations that proved to 190

204 be a successful way of avoiding dead-lock: when positions of different actors seemed irreconcilable, agreement was reached that with regard to a particular issue, there was a need for further research before definite decisions could be taken. This way a hopeless battle over who should get how much money in subsidies could be sidestepped (Canzler & Knie, 2011, p. 89 ff.). The need to find new ways of cooperating, in order to secure governmental support, further became clear to the members of the NPE when the government program for electric mobility was launched in 2011, because it did not hold out the prospect of funds for particular sectors and their individual interests but rather emphasized the need to strive for overarching goals (Canzler & Knie, 2011, p. 14). Outside these official and highly institutionalized groups, such as the NPE, especially the environmental and consumer associations offered a more critical perspective on e- mobility. They emphasized the need for a more comprehensive view on sustainability, thus broadening the dominant focus on electric vehicles to include issues of renewable energy sources for powering BEVs and visions of future more sustainable mobility in general. However, since they were not represented in the NPE their impact remained limited. Arguably, environmental associations, NGOs and independent environmental research institutes willingly positioned themselves outside the debate about e-mobility by taking an opposing stance even though the basic critique is less about e-mobility in general, but rather about the car-centered perspective dominating the political debate and initiatives (Zimmer & Rammler, 2011, p. 183 f.). In the overall political discourse, e-mobility appears as a promising option for mobility in the future that is currently struggling with technological problems such as limited range, long charging times and high costs, which are typical for an innovation in a premature state. These problems are to be overcome by fostering R&D and stimulating (technological) innovation. However, the fact that the problems dominating the debate are those that are most obvious when looking at the BEV in comparison to conventional cars (more expensive, smaller range, longer charging) rather than problems related to alternative forms of future mobility (transition to renewable energies, social aspects going beyond vehicle technology) shows that the current (auto)mobility regime remains the decisive frame of reference also with regard to visions of future mobility and the transformation process leading there. This is in line with the actor structure dominating the innovation system, i.e. representatives of OEMs, automotive suppliers and the energy sector, which make up the NPE and other influential bodies (Zimmer & Rammler, 2011, p. 184 f.). 191

205 Funding Programs: Trying to Achieve Systemic Change through Technological Innovation The most visible efforts in fostering e-mobility so far have been the three major funding programs Electric Mobility in Model Regions, Technological Lighthouses of Electric Mobility and Electric Mobility Showcases. The funding program Electric Mobility in Model Regions has been launched by the Federal Ministry of Transport, Building and Urban Development (BMVBS, now BMVI) and is coordinated by the National Organization for Hydrogen and Fuel Cell Technology (NOW GmbH). Eight regions have been selected as model regions for the period from 2009 to They were conceptualized as innovation clusters focusing on the following goals: technologyneutral R&D in the field of battery-electric vehicles, user-oriented demonstration, integration into transport, land use and urban planning and development, local integration of stakeholders and exchange of information across the model regions (Tenkhoff et al., 2011, p. 11). It can be shown how dominant rules and guiding principles manifest themselves in the way that financial resources have been allocated within the program. The volume of funding made available in this program amounted to 130 million from the Economic Stimulus Package II and thus made up a large part of the overall funding for electric mobility of 500 million in total. Roughly 70% of the model region funds were allocated to private businesses and together with industry investments the total project volume amounted to roughly 300 million. Across the model regions, 220 project partners were involved, 150 of which can be grouped among OEMs, component suppliers, energy suppliers and transport and logistics. In total, 70 individual demonstration projects focused on aspects of individual, commercial or public transport and in these projects approximately 2,500 electric vehicles (ranging from E-bikes to BEVs and commercial vehicles) were deployed and a charging infrastructure comprising roughly 1,000 charging stations was built up. Apart from technology-oriented research on vehicle technology, research was conducted on environmental effects, especially with regard to questions of available electricity sources, and on user acceptance of electric vehicles (Tenkhoff et al., 2011, p. 12). 192

206 Fig. 12: Electric Mobility in Model Regions Overview (Sax-Mobility II, 2012). The Federal Ministry of Education and Research (BMBF) has published a comprehensive overview of the government s funding activities in the field of e-mobility (BMBF, 2011). This overview contains lists of each individual amount that has been spent in the context of the Economic Stimulus Package II, in order to foster e-mobility. Based on this overview and the official evaluation report of the model regions (Tenkhoff et al., 2011), an overview has been gained on the number of individual projects within each of the eight model regions and the amount of money that has been allocated to these projects from the governmental funds. Then those projects have been selected that qualify as sustainable or system innovative according to the criteria deducted from sustainable mobility literatures (integration of renewable energy, fleet applications, intermodality, carsharing, see ch. 4.2). Any project that fulfills one or more of these criteria has been selected, based on the project description in the official evaluation report of the model region program. The share of funds that has been allocated to these projects has then been compared to the overall share a model region received for all projects. In the model region Rhein-Ruhr, 2 out of 8 projects focus on system-innovative elements, such as fleet applications and carsharing. In Stuttgart, 2 out of 7 projects focused on 193

207 fleet applications and (car-)sharing. In Bremen/Oldenburg, 3 out of 10 projects focused on carsharing. In Munich, 2 out of 4 projects focused on fleet applications and carsharing. In Saxony, 1 out of 3 projects focused on fleet applications. In Rhein-Main, 3 out of 16 projects focused on intermodality and carsharing. In Berlin/Potsdam, 2 out of 4 projects focused on intermodality and carsharing and on environmental effects of BEVs. In Hamburg, 1 out of 4 projects focused on the integration of renewable energies and fleet applications. The results of this quantitative comparison with regard to the share of funding allocated to these individual projects as related to the total amount of funding that a respective model region received is presented in Tbl. 5. Tbl. 5: Project funding in model regions and share of funding for system-innovative projects (data based on BMBF 2011, Tenkhoff et al., 2011). This overview shows that the amount of funding that goes to projects, which focus on systeminnovative aspects, is in most cases low to moderate. In two model regions, the share of funding allocated to system-innovative projects is higher than 50%: in Munich, it is 59.8% and in Berlin/Potsdam, it is 74.7%. In the case of Munich, a closer look at the individual projects shows that the 59.8% identified as funding for system-innovative projects, contain 54.5% going directly to Audi, a dominant regime actor. In the case of Berlin, 74.7% of the funding going to projects focusing on system-innovative aspects can be explained by the fact that in Berlin none of the large German car manufacturers are situated and as the capital it is a typical urban environment where public transport plays an important role. This is also the reason why for the case study Baden-Württemberg/Stuttgart have been chosen, because it is a more traditional automotive location with a large OEM, as well as a large energy supplier, while system-innovative approaches do still emerge to a relevant degree (apart from that efforts are continued in the follow-up showcase region program and a broad network of initiatives is developed, in contrast to some of the other model regions). In sum, this overview 194

208 shows that there is a willingness to invest large sums in e-mobility, while the projects that are chosen as eligible for funding remain largely within regime limits. However, according to the official evaluation report, the model regions have been successful and achieved the envisaged goals. Consequently, the Federal Ministry of Transport, Building and Urban Development (BMVBS, now BMVI) continues to follow this type of funding strategy, which centers around user-oriented demonstration within regional clusters, high industry participation and a technology-neutral approach (Tenkhoff et al., 2011, p. 31). This means that some of the model region projects are funded beyond the end of the official program in 2011 and a new follow-up funding program has been launched in 2012: the Electric Mobility Showcases (Bundesregierung, 2011, p. 28). In the description of the showcase program, signs can be found that there is more system-innovative potential: a systemic approach across the interface of the electric vehicle with the energy and the transport system has been chosen and a critical mass of vehicles is to be achieved, in order to learn about the suitability of electric mobility solutions (Bundesregierung, 2011, p. 26 f.). In April 2012, four showcase regions have been selected: Berlin/Brandenburg, Lower- Saxony, Baden-Württemberg and Bavaria/Saxony. The total volume of funding for the showcase region program is 180 million and each showcase will approximately receive up to 50 million in a period from 2013 to Fig. 13: Showcase Regions for Electric Mobility Overview (GGEMO, 2013). Four ministries are funding this program, the Federal Ministry of Economic and Innovation (BMWi) and the former Federal Ministry of Transport, Building and Urban 195